## What changes were proposed in this pull request?
This patch bumps master branch version to 2.2.0-SNAPSHOT.
## How was this patch tested?
N/A
Author: Reynold Xin <rxin@databricks.com>
Closes#16126 from rxin/SPARK-18695.
## What changes were proposed in this pull request?
SPARK-18429 introduced count-min sketch aggregate function for SQL, but the implementation and testing is more complicated than needed. This simplifies the test cases and removes support for data types that don't have clear equality semantics:
1. Removed support for floating point and decimal types.
2. Removed the heavy randomized tests. The underlying CountMinSketch implementation already had pretty good test coverage through randomized tests, and the SPARK-18429 implementation is just to add an aggregate function wrapper around CountMinSketch. There is no need for randomized tests at three different levels of the implementations.
## How was this patch tested?
A lot of the change is to simplify test cases.
Author: Reynold Xin <rxin@databricks.com>
Closes#16093 from rxin/SPARK-18663.
This PR separates the status of a `StreamingQuery` into two separate APIs:
- `status` - describes the status of a `StreamingQuery` at this moment, including what phase of processing is currently happening and if data is available.
- `recentProgress` - an array of statistics about the most recent microbatches that have executed.
A recent progress contains the following information:
```
{
"id" : "2be8670a-fce1-4859-a530-748f29553bb6",
"name" : "query-29",
"timestamp" : 1479705392724,
"inputRowsPerSecond" : 230.76923076923077,
"processedRowsPerSecond" : 10.869565217391303,
"durationMs" : {
"triggerExecution" : 276,
"queryPlanning" : 3,
"getBatch" : 5,
"getOffset" : 3,
"addBatch" : 234,
"walCommit" : 30
},
"currentWatermark" : 0,
"stateOperators" : [ ],
"sources" : [ {
"description" : "KafkaSource[Subscribe[topic-14]]",
"startOffset" : {
"topic-14" : {
"2" : 0,
"4" : 1,
"1" : 0,
"3" : 0,
"0" : 0
}
},
"endOffset" : {
"topic-14" : {
"2" : 1,
"4" : 2,
"1" : 0,
"3" : 0,
"0" : 1
}
},
"numRecords" : 3,
"inputRowsPerSecond" : 230.76923076923077,
"processedRowsPerSecond" : 10.869565217391303
} ]
}
```
Additionally, in order to make it possible to correlate progress updates across restarts, we change the `id` field from an integer that is unique with in the JVM to a `UUID` that is globally unique.
Author: Tathagata Das <tathagata.das1565@gmail.com>
Author: Michael Armbrust <michael@databricks.com>
Closes#15954 from marmbrus/queryProgress.
## What changes were proposed in this pull request?
Remove deprecated methods for ML.
## How was this patch tested?
Existing tests.
Author: Yanbo Liang <ybliang8@gmail.com>
Closes#15913 from yanboliang/spark-18481.
## What changes were proposed in this pull request?
When profiling heap dumps from the HistoryServer and live Spark web UIs, I found a large amount of memory being wasted on duplicated objects and strings. This patch's changes remove most of this duplication, resulting in over 40% memory savings for some benchmarks.
- **Task metrics** (6441f0624dfcda9c7193a64bfb416a145b5aabdf): previously, every `TaskUIData` object would have its own instances of `InputMetricsUIData`, `OutputMetricsUIData`, `ShuffleReadMetrics`, and `ShuffleWriteMetrics`, but for many tasks these metrics are irrelevant because they're all zero. This patch changes how we construct these metrics in order to re-use a single immutable "empty" value for the cases where these metrics are empty.
- **TaskInfo.accumulables** (ade86db901127bf13c0e0bdc3f09c933a093bb76): Previously, every `TaskInfo` object had its own empty `ListBuffer` for holding updates from named accumulators. Tasks which didn't use named accumulators still paid for the cost of allocating and storing this empty buffer. To avoid this overhead, I changed the `val` with a mutable buffer into a `var` which holds an immutable Scala list, allowing tasks which do not have named accumulator updates to share the same singleton `Nil` object.
- **String.intern() in JSONProtocol** (7e05630e9a78c455db8c8c499f0590c864624e05): in the HistoryServer, executor hostnames and ids are deserialized from JSON, leading to massive duplication of these string objects. By calling `String.intern()` on the deserialized values we can remove all of this duplication. Since Spark now requires Java 7+ we don't have to worry about string interning exhausting the permgen (see http://java-performance.info/string-intern-in-java-6-7-8/).
## How was this patch tested?
I ran
```
sc.parallelize(1 to 100000, 100000).count()
```
in `spark-shell` with event logging enabled, then loaded that event log in the HistoryServer, performed a full GC, and took a heap dump. According to YourKit, the changes in this patch reduced memory consumption by roughly 28 megabytes (or 770k Java objects):
Here's a table illustrating the drop in objects due to deduplication (the drop is <100k for some objects because some events were dropped from the listener bus; this is a separate, existing bug that I'll address separately after CPU-profiling):
Author: Josh Rosen <joshrosen@databricks.com>
Closes#15743 from JoshRosen/spark-ui-memory-usage.
We should upgrade to the latest release of MiMa (0.1.11) in order to include a fix for a bug which led to flakiness in the MiMa checks (https://github.com/typesafehub/migration-manager/issues/115).
Author: Josh Rosen <joshrosen@databricks.com>
Closes#15571 from JoshRosen/SPARK-18034.
## What changes were proposed in this pull request?
As per rxin request, here are further API changes
- Changed `Stream(Started/Progress/Terminated)` events to `Stream*Event`
- Changed the fields in `StreamingQueryListener.on***` from `query*` to `event`
## How was this patch tested?
Existing unit tests.
Author: Tathagata Das <tathagata.das1565@gmail.com>
Closes#15530 from tdas/SPARK-17731-1.
## What changes were proposed in this pull request?
Metrics are needed for monitoring structured streaming apps. Here is the design doc for implementing the necessary metrics.
https://docs.google.com/document/d/1NIdcGuR1B3WIe8t7VxLrt58TJB4DtipWEbj5I_mzJys/edit?usp=sharing
Specifically, this PR adds the following public APIs changes.
### New APIs
- `StreamingQuery.status` returns a `StreamingQueryStatus` object (renamed from `StreamingQueryInfo`, see later)
- `StreamingQueryStatus` has the following important fields
- inputRate - Current rate (rows/sec) at which data is being generated by all the sources
- processingRate - Current rate (rows/sec) at which the query is processing data from
all the sources
- ~~outputRate~~ - *Does not work with wholestage codegen*
- latency - Current average latency between the data being available in source and the sink writing the corresponding output
- sourceStatuses: Array[SourceStatus] - Current statuses of the sources
- sinkStatus: SinkStatus - Current status of the sink
- triggerStatus - Low-level detailed status of the last completed/currently active trigger
- latencies - getOffset, getBatch, full trigger, wal writes
- timestamps - trigger start, finish, after getOffset, after getBatch
- numRows - input, output, state total/updated rows for aggregations
- `SourceStatus` has the following important fields
- inputRate - Current rate (rows/sec) at which data is being generated by the source
- processingRate - Current rate (rows/sec) at which the query is processing data from the source
- triggerStatus - Low-level detailed status of the last completed/currently active trigger
- Python API for `StreamingQuery.status()`
### Breaking changes to existing APIs
**Existing direct public facing APIs**
- Deprecated direct public-facing APIs `StreamingQuery.sourceStatuses` and `StreamingQuery.sinkStatus` in favour of `StreamingQuery.status.sourceStatuses/sinkStatus`.
- Branch 2.0 should have it deprecated, master should have it removed.
**Existing advanced listener APIs**
- `StreamingQueryInfo` renamed to `StreamingQueryStatus` for consistency with `SourceStatus`, `SinkStatus`
- Earlier StreamingQueryInfo was used only in the advanced listener API, but now it is used in direct public-facing API (StreamingQuery.status)
- Field `queryInfo` in listener events `QueryStarted`, `QueryProgress`, `QueryTerminated` changed have name `queryStatus` and return type `StreamingQueryStatus`.
- Field `offsetDesc` in `SourceStatus` was Option[String], converted it to `String`.
- For `SourceStatus` and `SinkStatus` made constructor private instead of private[sql] to make them more java-safe. Instead added `private[sql] object SourceStatus/SinkStatus.apply()` which are harder to accidentally use in Java.
## How was this patch tested?
Old and new unit tests.
- Rate calculation and other internal logic of StreamMetrics tested by StreamMetricsSuite.
- New info in statuses returned through StreamingQueryListener is tested in StreamingQueryListenerSuite.
- New and old info returned through StreamingQuery.status is tested in StreamingQuerySuite.
- Source-specific tests for making sure input rows are counted are is source-specific test suites.
- Additional tests to test minor additions in LocalTableScanExec, StateStore, etc.
Metrics also manually tested using Ganglia sink
Author: Tathagata Das <tathagata.das1565@gmail.com>
Closes#15307 from tdas/SPARK-17731.
## What changes were proposed in this pull request?
address post hoc review comments for https://github.com/apache/spark/pull/14897
## How was this patch tested?
N/A
Author: Wenchen Fan <wenchen@databricks.com>
Closes#15424 from cloud-fan/global-temp-view.
## What changes were proposed in this pull request?
Global temporary view is a cross-session temporary view, which means it's shared among all sessions. Its lifetime is the lifetime of the Spark application, i.e. it will be automatically dropped when the application terminates. It's tied to a system preserved database `global_temp`(configurable via SparkConf), and we must use the qualified name to refer a global temp view, e.g. SELECT * FROM global_temp.view1.
changes for `SessionCatalog`:
1. add a new field `gloabalTempViews: GlobalTempViewManager`, to access the shared global temp views, and the global temp db name.
2. `createDatabase` will fail if users wanna create `global_temp`, which is system preserved.
3. `setCurrentDatabase` will fail if users wanna set `global_temp`, which is system preserved.
4. add `createGlobalTempView`, which is used in `CreateViewCommand` to create global temp views.
5. add `dropGlobalTempView`, which is used in `CatalogImpl` to drop global temp view.
6. add `alterTempViewDefinition`, which is used in `AlterViewAsCommand` to update the view definition for local/global temp views.
7. `renameTable`/`dropTable`/`isTemporaryTable`/`lookupRelation`/`getTempViewOrPermanentTableMetadata`/`refreshTable` will handle global temp views.
changes for SQL commands:
1. `CreateViewCommand`/`AlterViewAsCommand` is updated to support global temp views
2. `ShowTablesCommand` outputs a new column `database`, which is used to distinguish global and local temp views.
3. other commands can also handle global temp views if they call `SessionCatalog` APIs which accepts global temp views, e.g. `DropTableCommand`, `AlterTableRenameCommand`, `ShowColumnsCommand`, etc.
changes for other public API
1. add a new method `dropGlobalTempView` in `Catalog`
2. `Catalog.findTable` can find global temp view
3. add a new method `createGlobalTempView` in `Dataset`
## How was this patch tested?
new tests in `SQLViewSuite`
Author: Wenchen Fan <wenchen@databricks.com>
Closes#14897 from cloud-fan/global-temp-view.
## What changes were proposed in this pull request?
Return Iterator of applications internally in history server, for consistency and performance. See https://github.com/apache/spark/pull/15248 for some back-story.
The code called by and calling HistoryServer.getApplicationList wants an Iterator, but this method materializes an Iterable, which potentially causes a performance problem. It's simpler too to make this internal method also pass through an Iterator.
## How was this patch tested?
Existing tests.
Author: Sean Owen <sowen@cloudera.com>
Closes#15321 from srowen/SPARK-17671.
## What changes were proposed in this pull request?
We added find and exists methods for Databases, Tables and Functions to the user facing Catalog in PR https://github.com/apache/spark/pull/15301. However, it was brought up that the semantics of the `find` methods are more in line a `get` method (get an object or else fail). So we rename these in this PR.
## How was this patch tested?
Existing tests.
Author: Herman van Hovell <hvanhovell@databricks.com>
Closes#15308 from hvanhovell/SPARK-17717-2.
## What changes were proposed in this pull request?
The current user facing catalog does not implement methods for checking object existence or finding objects. You could theoretically do this using the `list*` commands, but this is rather cumbersome and can actually be costly when there are many objects. This PR adds `exists*` and `find*` methods for Databases, Table and Functions.
## How was this patch tested?
Added tests to `org.apache.spark.sql.internal.CatalogSuite`
Author: Herman van Hovell <hvanhovell@databricks.com>
Closes#15301 from hvanhovell/SPARK-17717.
## What changes were proposed in this pull request?
Several performance improvement for ```ChiSqSelector```:
1, Keep ```selectedFeatures``` ordered ascendent.
```ChiSqSelectorModel.transform``` need ```selectedFeatures``` ordered to make prediction. We should sort it when training model rather than making prediction, since users usually train model once and use the model to do prediction multiple times.
2, When training ```fpr``` type ```ChiSqSelectorModel```, it's not necessary to sort the ChiSq test result by statistic.
## How was this patch tested?
Existing unit tests.
Author: Yanbo Liang <ybliang8@gmail.com>
Closes#15277 from yanboliang/spark-17704.
Currently task metrics don't support executor CPU time, so there's no way to calculate how much CPU time a stage/task took from History Server metrics. This PR enables reporting CPU time.
Author: jisookim <jisookim0513@gmail.com>
Closes#10212 from jisookim0513/add-cpu-time-metric.
## What changes were proposed in this pull request?
Allow Spark 2.x to load instances of LDA, LocalLDAModel, and DistributedLDAModel saved from Spark 1.6.
## How was this patch tested?
I tested this manually, saving the 3 types from 1.6 and loading them into master (2.x). In the future, we can add generic tests for testing backwards compatibility across all ML models in SPARK-15573.
Author: Joseph K. Bradley <joseph@databricks.com>
Closes#15034 from jkbradley/lda-backwards.
## What changes were proposed in this pull request?
We are killing multiple executors together instead of iterating over expensive RPC calls to kill single executor.
## How was this patch tested?
Executed sample spark job to observe executors being killed/removed with dynamic allocation enabled.
Author: Dhruve Ashar <dashar@yahoo-inc.com>
Author: Dhruve Ashar <dhruveashar@gmail.com>
Closes#15152 from dhruve/impr/SPARK-17365.
## What changes were proposed in this pull request?
Univariate feature selection works by selecting the best features based on univariate statistical tests. False Positive Rate (FPR) is a popular univariate statistical test for feature selection. We add a chiSquare Selector based on False Positive Rate (FPR) test in this PR, like it is implemented in scikit-learn.
http://scikit-learn.org/stable/modules/feature_selection.html#univariate-feature-selection
## How was this patch tested?
Add Scala ut
Author: Peng, Meng <peng.meng@intel.com>
Closes#14597 from mpjlu/fprChiSquare.
## What changes were proposed in this pull request?
Merge `MultinomialLogisticRegression` into `LogisticRegression` and remove `MultinomialLogisticRegression`.
Marked as WIP because we should discuss the coefficients API in the model. See discussion below.
JIRA: [SPARK-17163](https://issues.apache.org/jira/browse/SPARK-17163)
## How was this patch tested?
Merged test suites and added some new unit tests.
## Design
### Switching between binomial and multinomial
We default to automatically detecting whether we should run binomial or multinomial lor. We expose a new parameter called `family` which defaults to auto. When "auto" is used, we run normal binomial lor with pivoting if there are 1 or 2 label classes. Otherwise, we run multinomial. If the user explicitly sets the family, then we abide by that setting. In the case where "binomial" is set but multiclass lor is detected, we throw an error.
### coefficients/intercept model API (TODO)
This is the biggest design point remaining, IMO. We need to decide how to store the coefficients and intercepts in the model, and in turn how to expose them via the API. Two important points:
* We must maintain compatibility with the old API, i.e. we must expose `def coefficients: Vector` and `def intercept: Double`
* There are two separate cases: binomial lr where we have a single set of coefficients and a single intercept and multinomial lr where we have `numClasses` sets of coefficients and `numClasses` intercepts.
Some options:
1. **Store the binomial coefficients as a `2 x numFeatures` matrix.** This means that we would center the model coefficients before storing them in the model. The BLOR algorithm gives `1 * numFeatures` coefficients, but we would convert them to `2 x numFeatures` coefficients before storing them, effectively doubling the storage in the model. This has the advantage that we can make the code cleaner (i.e. less `if (isMultinomial) ... else ...`) and we don't have to reason about the different cases as much. It has the disadvantage that we double the storage space and we could see small regressions at prediction time since there are 2x the number of operations in the prediction algorithms. Additionally, we still have to produce the uncentered coefficients/intercept via the API, so we will have to either ALSO store the uncentered version, or compute it in `def coefficients: Vector` every time.
2. **Store the binomial coefficients as a `1 x numFeatures` matrix.** We still store the coefficients as a matrix and the intercepts as a vector. When users call `coefficients` we return them a `Vector` that is backed by the same underlying array as the `coefficientMatrix`, so we don't duplicate any data. At prediction time, we use the old prediction methods that are specialized for binary LOR. The benefits here are that we don't store extra data, and we won't see any regressions in performance. The cost of this is that we have separate implementations for predict methods in the binary vs multiclass case. The duplicated code is really not very high, but it's still a bit messy.
If we do decide to store the 2x coefficients, we would likely want to see some performance tests to understand the potential regressions.
**Update:** We have chosen option 2
### Threshold/thresholds (TODO)
Currently, when `threshold` is set we clear whatever value is in `thresholds` and when `thresholds` is set we clear whatever value is in `threshold`. [SPARK-11543](https://issues.apache.org/jira/browse/SPARK-11543) was created to prefer thresholds over threshold. We should decide if we should implement this behavior now or if we want to do it in a separate JIRA.
**Update:** Let's leave it for a follow up PR
## Follow up
* Summary model for multiclass logistic regression [SPARK-17139](https://issues.apache.org/jira/browse/SPARK-17139)
* Thresholds vs threshold [SPARK-11543](https://issues.apache.org/jira/browse/SPARK-11543)
Author: sethah <seth.hendrickson16@gmail.com>
Closes#14834 from sethah/SPARK-17163.
## What changes were proposed in this pull request?
Following https://github.com/apache/spark/pull/14969 for some reason the MiMa excludes weren't complete, but still passed the PR builder. This adds 3 more excludes from https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-sbt-hadoop-2.2/1749/consoleFull
It also moves the excludes to their own Seq in the build, as they probably should have been.
Even though this is merged to 2.1.x only / master, I left the exclude in for 2.0.x in case we back port. It's a private API so is always a false positive.
## How was this patch tested?
Jenkins build
Author: Sean Owen <sowen@cloudera.com>
Closes#15110 from srowen/SPARK-17406.2.
## What changes were proposed in this pull request?
The job page will be too slow to open when there are thousands of executor events(added or removed). I found that in ExecutorsTab file, executorIdToData will not remove elements, it will increase all the time.Before this pr, it looks like [timeline1.png](https://issues.apache.org/jira/secure/attachment/12827112/timeline1.png). After this pr, it looks like [timeline2.png](https://issues.apache.org/jira/secure/attachment/12827113/timeline2.png)(we can set how many executor events will be displayed)
Author: cenyuhai <cenyuhai@didichuxing.com>
Closes#14969 from cenyuhai/SPARK-17406.
This patch makes a handful of post-Spark-2.0 MiMa exclusion and build updates. It should be merged to master and a subset of it should be picked into branch-2.0 in order to test Spark 2.0.1-SNAPSHOT.
- Remove the ` sketch`, `mllibLocal`, and `streamingKafka010` from the list of excluded subprojects so that MiMa checks them.
- Remove now-unnecessary special-case handling of the Kafka 0.8 artifact in `mimaSettings`.
- Move the exclusion added in SPARK-14743 from `v20excludes` to `v21excludes`, since that patch was only merged into master and not branch-2.0.
- Add exclusions for an API change introduced by SPARK-17096 / #14675.
- Add missing exclusions for the `o.a.spark.internal` and `o.a.spark.sql.internal` packages.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#15061 from JoshRosen/post-2.0-mima-changes.
## What changes were proposed in this pull request?
Move Mesos code into a mvn module
## How was this patch tested?
unit tests
manually submitting a client mode and cluster mode job
spark/mesos integration test suite
Author: Michael Gummelt <mgummelt@mesosphere.io>
Closes#14637 from mgummelt/mesos-module.
## What changes were proposed in this pull request?
Add a configurable token manager for Spark on running on yarn.
### Current Problems ###
1. Supported token provider is hard-coded, currently only hdfs, hbase and hive are supported and it is impossible for user to add new token provider without code changes.
2. Also this problem exits in timely token renewer and updater.
### Changes In This Proposal ###
In this proposal, to address the problems mentioned above and make the current code more cleaner and easier to understand, mainly has 3 changes:
1. Abstract a `ServiceTokenProvider` as well as `ServiceTokenRenewable` interface for token provider. Each service wants to communicate with Spark through token way needs to implement this interface.
2. Provide a `ConfigurableTokenManager` to manage all the register token providers, also token renewer and updater. Also this class offers the API for other modules to obtain tokens, get renewal interval and so on.
3. Implement 3 built-in token providers `HDFSTokenProvider`, `HiveTokenProvider` and `HBaseTokenProvider` to keep the same semantics as supported today. Whether to load in these built-in token providers is controlled by configuration "spark.yarn.security.tokens.${service}.enabled", by default for all the built-in token providers are loaded.
### Behavior Changes ###
For the end user there's no behavior change, we still use the same configuration `spark.yarn.security.tokens.${service}.enabled` to decide which token provider is enabled (hbase or hive).
For user implemented token provider (assume the name of token provider is "test") needs to add into this class should have two configurations:
1. `spark.yarn.security.tokens.test.enabled` to true
2. `spark.yarn.security.tokens.test.class` to the full qualified class name.
So we still keep the same semantics as current code while add one new configuration.
### Current Status ###
- [x] token provider interface and management framework.
- [x] implement built-in token providers (hdfs, hbase, hive).
- [x] Coverage of unit test.
- [x] Integrated test with security cluster.
## How was this patch tested?
Unit test and integrated test.
Please suggest and review, any comment is greatly appreciated.
Author: jerryshao <sshao@hortonworks.com>
Closes#14065 from jerryshao/SPARK-16342.
## What changes were proposed in this pull request?
For DataSet typed select:
```
def select[U1: Encoder](c1: TypedColumn[T, U1]): Dataset[U1]
```
If type T is a case class or a tuple class that is not atomic, the resulting logical plan's schema will mismatch with `Dataset[T]` encoder's schema, which will cause encoder error and throw AnalysisException.
### Before change:
```
scala> case class A(a: Int, b: Int)
scala> Seq((0, A(1,2))).toDS.select($"_2".as[A])
org.apache.spark.sql.AnalysisException: cannot resolve '`a`' given input columns: [_2];
..
```
### After change:
```
scala> case class A(a: Int, b: Int)
scala> Seq((0, A(1,2))).toDS.select($"_2".as[A]).show
+---+---+
| a| b|
+---+---+
| 1| 2|
+---+---+
```
## How was this patch tested?
Unit test.
Author: Sean Zhong <seanzhong@databricks.com>
Closes#14474 from clockfly/SPARK-16853.
## What changes were proposed in this pull request?
It would be useful to support listing the columns that are referenced by a filter. This can help simplify data source planning, because with this we would be able to implement unhandledFilters method in HadoopFsRelation.
This is based on rxin's patch (#13901) and adds unit tests.
## How was this patch tested?
Added a new suite FiltersSuite.
Author: petermaxlee <petermaxlee@gmail.com>
Author: Reynold Xin <rxin@databricks.com>
Closes#14120 from petermaxlee/SPARK-16199.
## What changes were proposed in this pull request?
It is currently fairly difficult to have proper mima excludes when we cut a version branch. I'm proposing a small change to take the exclude list out of the exclude function, and put it in a variable so we can easily union excludes.
After this change, we can bump pom.xml version to 2.1.0-SNAPSHOT, without bumping the diff base version. Note that I also deleted all the exclude rules for version 1.x, to cut down the size of the file.
## How was this patch tested?
N/A - this is a build infra change.
Author: Reynold Xin <rxin@databricks.com>
Closes#14128 from rxin/SPARK-16476.
## What changes were proposed in this pull request?
Revert partial changes in SPARK-12600, and add some deprecated method back to SQLContext for backward source code compatibility.
## How was this patch tested?
Manual test.
Author: Sean Zhong <seanzhong@databricks.com>
Closes#13637 from clockfly/SPARK-15914.
## What changes were proposed in this pull request?
We're using `asML` to convert the mllib vector/matrix to ml vector/matrix now. Using `as` is more correct given that this conversion actually shares the same underline data structure. As a result, in this PR, `toBreeze` will be changed to `asBreeze`. This is a private API, as a result, it will not affect any user's application.
## How was this patch tested?
unit tests
Author: DB Tsai <dbt@netflix.com>
Closes#13198 from dbtsai/minor.
## What changes were proposed in this pull request?
This PR changes SQLContext/HiveContext's public constructor to use SparkSession.build.getOrCreate and removes isRootContext from SQLContext.
## How was this patch tested?
Existing tests.
Author: Yin Huai <yhuai@databricks.com>
Closes#13310 from yhuai/SPARK-15532.
## What changes were proposed in this pull request?
This patch renames various DefaultSources to make their names more self-describing. The choice of "DefaultSource" was from the days when we did not have a good way to specify short names.
They are now named:
- LibSVMFileFormat
- CSVFileFormat
- JdbcRelationProvider
- JsonFileFormat
- ParquetFileFormat
- TextFileFormat
Backward compatibility is maintained through aliasing.
## How was this patch tested?
Updated relevant test cases too.
Author: Reynold Xin <rxin@databricks.com>
Closes#13311 from rxin/SPARK-15543.
## What changes were proposed in this pull request?
Since we support forced spilling for Spillable, which only works in OnHeap mode, different from other SQL operators (could be OnHeap or OffHeap), we should considering the mode of consumer before calling trigger forced spilling.
## How was this patch tested?
Add new test.
Author: Davies Liu <davies@databricks.com>
Closes#13151 from davies/fix_mode.
## What changes were proposed in this pull request?
Once SPARK-14487 and SPARK-14549 are merged, we will migrate to use the new vector and matrix type in the new ml pipeline based apis.
## How was this patch tested?
Unit tests
Author: DB Tsai <dbt@netflix.com>
Author: Liang-Chi Hsieh <simonh@tw.ibm.com>
Author: Xiangrui Meng <meng@databricks.com>
Closes#12627 from dbtsai/SPARK-14615-NewML.
## What changes were proposed in this pull request?
(See https://github.com/apache/spark/pull/12416 where most of this was already reviewed and committed; this is just the module structure and move part. This change does not move the annotations into test scope, which was the apparently problem last time.)
Rename `spark-test-tags` -> `spark-tags`; move common annotations like `Since` to `spark-tags`
## How was this patch tested?
Jenkins tests.
Author: Sean Owen <sowen@cloudera.com>
Closes#13074 from srowen/SPARK-15290.
## What changes were proposed in this pull request?
This PR removes the old `json(path: String)` API which is covered by the new `json(paths: String*)`.
## How was this patch tested?
Jenkins tests (existing tests should cover this)
Author: hyukjinkwon <gurwls223@gmail.com>
Author: Hyukjin Kwon <gurwls223@gmail.com>
Closes#13040 from HyukjinKwon/SPARK-15250.
## What changes were proposed in this pull request?
Currently PipedRDD internally uses PrintWriter to write data to the stdin of the piped process, which by default uses a BufferedWriter of buffer size 8k. In our experiment, we have seen that 8k buffer size is too small and the job spends significant amount of CPU time in system calls to copy the data. We should have a way to configure the buffer size for the writer.
## How was this patch tested?
Ran PipedRDDSuite tests.
Author: Sital Kedia <skedia@fb.com>
Closes#12309 from sitalkedia/bufferedPipedRDD.
## What changes were proposed in this pull request?
Removed blockTransferService and sparkFilesDir from SparkEnv since they're rarely used and don't need to be in stored in the env. Edited their few usages to accommodate the change.
## How was this patch tested?
ran dev/run-tests locally
Author: Alex Bozarth <ajbozart@us.ibm.com>
Closes#12970 from ajbozarth/spark10653.
#### What changes were proposed in this pull request?
This PR removes three methods the were deprecated in 1.6.0:
- `PortableDataStream.close()`
- `LinearRegression.weights`
- `LogisticRegression.weights`
The rationale for doing this is that the impact is small and that Spark 2.0 is a major release.
#### How was this patch tested?
Compilation succeded.
Author: Herman van Hovell <hvanhovell@questtec.nl>
Closes#12732 from hvanhovell/SPARK-14952.
## What changes were proposed in this pull request?
This PR adds `since` tag into the matrix and vector classes in spark-mllib-local.
## How was this patch tested?
Scala-style checks passed.
Author: Pravin Gadakh <prgadakh@in.ibm.com>
Closes#12416 from pravingadakh/SPARK-14613.
## What changes were proposed in this pull request?
This PR introduces a new accumulator API which is much simpler than before:
1. the type hierarchy is simplified, now we only have an `Accumulator` class
2. Combine `initialValue` and `zeroValue` concepts into just one concept: `zeroValue`
3. there in only one `register` method, the accumulator registration and cleanup registration are combined.
4. the `id`,`name` and `countFailedValues` are combined into an `AccumulatorMetadata`, and is provided during registration.
`SQLMetric` is a good example to show the simplicity of this new API.
What we break:
1. no `setValue` anymore. In the new API, the intermedia type can be different from the result type, it's very hard to implement a general `setValue`
2. accumulator can't be serialized before registered.
Problems need to be addressed in follow-ups:
1. with this new API, `AccumulatorInfo` doesn't make a lot of sense, the partial output is not partial updates, we need to expose the intermediate value.
2. `ExceptionFailure` should not carry the accumulator updates. Why do users care about accumulator updates for failed cases? It looks like we only use this feature to update the internal metrics, how about we sending a heartbeat to update internal metrics after the failure event?
3. the public event `SparkListenerTaskEnd` carries a `TaskMetrics`. Ideally this `TaskMetrics` don't need to carry external accumulators, as the only method of `TaskMetrics` that can access external accumulators is `private[spark]`. However, `SQLListener` use it to retrieve sql metrics.
## How was this patch tested?
existing tests
Author: Wenchen Fan <wenchen@databricks.com>
Closes#12612 from cloud-fan/acc.
## What changes were proposed in this pull request?
In Spark 2.0, `SparkSession` is the new thing. Internally we should stop using `SQLContext` everywhere since that's supposed to be not the main user-facing API anymore.
In this patch I took care to not break any public APIs. The one place that's suspect is `o.a.s.ml.source.libsvm.DefaultSource`, but according to mengxr it's not supposed to be public so it's OK to change the underlying `FileFormat` trait.
**Reviewers**: This is a big patch that may be difficult to review but the changes are actually really straightforward. If you prefer I can break it up into a few smaller patches, but it will delay the progress of this issue a little.
## How was this patch tested?
No change in functionality intended.
Author: Andrew Or <andrew@databricks.com>
Closes#12625 from andrewor14/spark-session-refactor.
## What changes were proposed in this pull request?
Implement some `hashCode` and `equals` together in order to enable the scalastyle.
This is a first batch, I will continue to implement them but I wanted to know your thoughts.
Author: Joan <joan@goyeau.com>
Closes#12157 from joan38/SPARK-6429-HashCode-Equals.
## What changes were proposed in this pull request?
For maintaining wrappers around spark.mllib algorithms in spark.ml, it will be useful to have ```private[spark]``` methods for converting from one linear algebra representation to another.
This PR adds toNew, fromNew methods for all spark.mllib Vector and Matrix types.
## How was this patch tested?
Unit tests for all conversions
Author: Joseph K. Bradley <joseph@databricks.com>
Closes#12504 from jkbradley/linalg-conversions.
## What changes were proposed in this pull request?
After removing most of `HiveContext` in 8fc267ab33 we can now move existing functionality in `SQLContext` to `SparkSession`. As of this PR `SQLContext` becomes a simple wrapper that has a `SparkSession` and delegates all functionality to it.
## How was this patch tested?
Jenkins.
Author: Andrew Or <andrew@databricks.com>
Closes#12553 from andrewor14/implement-spark-session.
## What changes were proposed in this pull request?
In #9241 It implemented a mechanism to call spill() on those SQL operators that support spilling if there is not enough memory for execution.
But ExternalSorter and AppendOnlyMap in Spark core are not worked. So this PR make them benefit from #9241. Now when there is not enough memory for execution, it can get memory by spilling ExternalSorter and AppendOnlyMap in Spark core.
## How was this patch tested?
add two unit tests for it.
Author: Lianhui Wang <lianhuiwang09@gmail.com>
Closes#10024 from lianhuiwang/SPARK-4452-2.
## What changes were proposed in this pull request?
Before this PR, we create accumulators at driver side(and register them) and send them to executor side, then we create `TaskMetrics` with these accumulators at executor side.
After this PR, we will create `TaskMetrics` at driver side and send it to executor side, so that we can create accumulators inside `TaskMetrics` directly, which is cleaner.
## How was this patch tested?
existing tests.
Author: Wenchen Fan <wenchen@databricks.com>
Closes#12472 from cloud-fan/acc.
## What changes were proposed in this pull request?
This PR moves `HadoopFsRelation` related data source API into `execution/datasources` package.
Note that to avoid conflicts, this PR is based on #12153. Effective changes for this PR only consist of the last three commits. Will rebase after merging #12153.
## How was this patch tested?
Existing tests.
Author: Yin Huai <yhuai@databricks.com>
Author: Cheng Lian <lian@databricks.com>
Closes#12361 from liancheng/spark-14407-hide-hadoop-fs-relation.
## What changes were proposed in this pull request?
This PR adds support for specifying an optional custom coalescer to the `coalesce()` method. Currently I have only added this feature to the `RDD` interface, and once we sort out the details we can proceed with adding this feature to the other APIs (`Dataset` etc.)
## How was this patch tested?
Added a unit test for this functionality.
/cc rxin (per our discussion on the mailing list)
Author: Nezih Yigitbasi <nyigitbasi@netflix.com>
Closes#11865 from nezihyigitbasi/custom_coalesce_policy.
## What changes were proposed in this pull request?
This PR is a follow up for https://github.com/apache/spark/pull/12417, now we always track input/output/shuffle metrics in spark JSON protocol and status API.
Most of the line changes are because of re-generating the gold answer for `HistoryServerSuite`, and we add a lot of 0 values for read/write metrics.
## How was this patch tested?
existing tests.
Author: Wenchen Fan <wenchen@databricks.com>
Closes#12462 from cloud-fan/follow.
## What changes were proposed in this pull request?
Part of the reason why TaskMetrics and its callers are complicated are due to the optional metrics we collect, including input, output, shuffle read, and shuffle write. I think we can always track them and just assign 0 as the initial values. It is usually very obvious whether a task is supposed to read any data or not. By always tracking them, we can remove a lot of map, foreach, flatMap, getOrElse(0L) calls throughout Spark.
This patch also changes a few behaviors.
1. Removed the distinction of data read/write methods (e.g. Hadoop, Memory, Network, etc).
2. Accumulate all data reads and writes, rather than only the first method. (Fixes SPARK-5225)
## How was this patch tested?
existing tests.
This is bases on https://github.com/apache/spark/pull/12388, with more test fixes.
Author: Reynold Xin <rxin@databricks.com>
Author: Wenchen Fan <wenchen@databricks.com>
Closes#12417 from cloud-fan/metrics-refactor.
## What changes were proposed in this pull request?
This patch removes some of the deprecated APIs in TaskMetrics. This is part of my bigger effort to simplify accumulators and task metrics.
## How was this patch tested?
N/A - only removals
Author: Reynold Xin <rxin@databricks.com>
Closes#12375 from rxin/SPARK-14617.
## What changes were proposed in this pull request?
Old `HadoopFsRelation` API includes `buildInternalScan()` which uses `SqlNewHadoopRDD` in `ParquetRelation`.
Because now the old API is removed, `SqlNewHadoopRDD` is not used anymore.
So, this PR removes `SqlNewHadoopRDD` and several unused imports.
This was discussed in https://github.com/apache/spark/pull/12326.
## How was this patch tested?
Several related existing unit tests and `sbt scalastyle`.
Author: hyukjinkwon <gurwls223@gmail.com>
Closes#12354 from HyukjinKwon/SPARK-14596.
## What changes were proposed in this pull request?
This adds a new API call `TaskContext.getLocalProperty` for getting properties set in the driver from executors. These local properties are automatically propagated from the driver to executors. For streaming, the context for streaming tasks will be the initial driver context when ssc.start() is called.
## How was this patch tested?
Unit tests.
cc JoshRosen
Author: Eric Liang <ekl@databricks.com>
Closes#12248 from ericl/sc-2813.
## What changes were proposed in this pull request?
When we first introduced Aggregators, we required the user of Aggregators to (implicitly) specify the encoders. It would actually make more sense to have the encoders be specified by the implementation of Aggregators, since each implementation should have the most state about how to encode its own data type.
Note that this simplifies the Java API because Java users no longer need to explicitly specify encoders for aggregators.
## How was this patch tested?
Updated unit tests.
Author: Reynold Xin <rxin@databricks.com>
Closes#12231 from rxin/SPARK-14451.
## What changes were proposed in this pull request?
Here is why SPARK-14437 happens:
BlockManagerId is created using NettyBlockTransferService.hostName which comes from `customHostname`. And `Executor` will set `customHostname` to the hostname which is detected by the driver. However, the driver may not be able to detect the correct address in some complicated network (Netty's Channel.remoteAddress doesn't always return a connectable address). In such case, `BlockManagerId` will be created using a wrong hostname.
To fix this issue, this PR uses `hostname` provided by `SparkEnv.create` to create `NettyBlockTransferService` and set `NettyBlockTransferService.hostname` to this one directly. A bonus of this approach is NettyBlockTransferService won't bound to `0.0.0.0` which is much safer.
## How was this patch tested?
Manually checked the bound address using local-cluster.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#12240 from zsxwing/SPARK-14437.
## What changes were proposed in this pull request?
The EMLDAOptimizer should generally not delete its last checkpoint since that can cause failures when DistributedLDAModel methods are called (if any partitions need to be recovered from the checkpoint).
This PR adds a "deleteLastCheckpoint" option which defaults to false. This is a change in behavior from Spark 1.6, in that the last checkpoint will not be removed by default.
This involves adding the deleteLastCheckpoint option to both spark.ml and spark.mllib, and modifying PeriodicCheckpointer to support the option.
This also:
* Makes MLlibTestSparkContext extend TempDirectory and set the checkpointDir to tempDir
* Updates LibSVMRelationSuite because of a name conflict with "tempDir" (and fixes a bug where it failed to delete a temp directory)
* Adds a MIMA exclude for DistributedLDAModel constructor, which is already ```private[clustering]```
## How was this patch tested?
Added 2 new unit tests to spark.ml LDASuite, which calls into spark.mllib.
Author: Joseph K. Bradley <joseph@databricks.com>
Closes#12166 from jkbradley/emlda-save-checkpoint.
Currently all `SparkFirehoseListener` implementations are broken since we expect listeners to extend `SparkListener`, while the fire hose only extends `SparkListenerInterface`. This changes the addListener function and the config based injection to use the interface instead.
The existing tests in SparkListenerSuite are improved such that they would have caught this.
Follow-up to #12142
Author: Michael Armbrust <michael@databricks.com>
Closes#12227 from marmbrus/fixListener.
## What changes were proposed in this pull request?
Adding Python API for training summaries of LogisticRegression and LinearRegression in PySpark ML.
## How was this patch tested?
Added unit tests to exercise the api calls for the summary classes. Also, manually verified values are expected and match those from Scala directly.
Author: Bryan Cutler <cutlerb@gmail.com>
Closes#11621 from BryanCutler/pyspark-ml-summary-SPARK-13430.
## What changes were proposed in this pull request?
Scala traits are difficult to maintain binary compatibility on, and as a result we had to introduce JavaSparkListener. In Spark 2.0 we can change SparkListener from a trait to an abstract class and then remove JavaSparkListener.
## How was this patch tested?
Updated related unit tests.
Author: Reynold Xin <rxin@databricks.com>
Closes#12142 from rxin/SPARK-14358.
JIRA: https://issues.apache.org/jira/browse/SPARK-13674
## What changes were proposed in this pull request?
Sample operator doesn't support wholestage codegen now. This pr is to add support to it.
## How was this patch tested?
A test is added into `BenchmarkWholeStageCodegen`. Besides, all tests should be passed.
Author: Liang-Chi Hsieh <simonh@tw.ibm.com>
Author: Liang-Chi Hsieh <viirya@gmail.com>
Closes#11517 from viirya/add-wholestage-sample.
1.Implement LossFunction trait and implement squared error and cross entropy
loss with it
2.Implement unit test for gradient and loss
3.Implement InPlace trait and in-place layer evaluation
4.Refactor interface for ActivationFunction
5.Update of Layer and LayerModel interfaces
6.Fix random weights assignment
7.Implement memory allocation by MLP model instead of individual layers
These features decreased the memory usage and increased flexibility of
internal API.
Author: Alexander Ulanov <nashb@yandex.ru>
Author: avulanov <avulanov@gmail.com>
Closes#9229 from avulanov/mlp-refactoring.
## What changes were proposed in this pull request?
After DataFrame and Dataset are merged, the trait `Queryable` becomes unnecessary as it has only one implementation. We should remove it.
## How was this patch tested?
existing tests.
Author: Wenchen Fan <wenchen@databricks.com>
Closes#12001 from cloud-fan/df-ds.
## What changes were proposed in this pull request?
Removed methods that has been deprecated since 1.1, 1.2, 1.3, 1.4, and 1.5.
## How was this patch tested?
- manully checked that no codes in Spark call these methods any more
- existing test suits
Author: Liwei Lin <lwlin7@gmail.com>
Author: proflin <proflin.me@gmail.com>
Closes#11910 from lw-lin/remove-deprecates.
## What changes were proposed in this pull request?
This reopens#11836, which was merged but promptly reverted because it introduced flaky Hive tests.
## How was this patch tested?
See `CatalogTestCases`, `SessionCatalogSuite` and `HiveContextSuite`.
Author: Andrew Or <andrew@databricks.com>
Closes#11938 from andrewor14/session-catalog-again.
## What changes were proposed in this pull request?
`SessionCatalog`, introduced in #11750, is a catalog that keeps track of temporary functions and tables, and delegates metastore operations to `ExternalCatalog`. This functionality overlaps a lot with the existing `analysis.Catalog`.
As of this commit, `SessionCatalog` and `ExternalCatalog` will no longer be dead code. There are still things that need to be done after this patch, namely:
- SPARK-14013: Properly implement temporary functions in `SessionCatalog`
- SPARK-13879: Decide which DDL/DML commands to support natively in Spark
- SPARK-?????: Implement the ones we do want to support through `SessionCatalog`.
- SPARK-?????: Merge SQL/HiveContext
## How was this patch tested?
This is largely a refactoring task so there are no new tests introduced. The particularly relevant tests are `SessionCatalogSuite` and `ExternalCatalogSuite`.
Author: Andrew Or <andrew@databricks.com>
Author: Yin Huai <yhuai@databricks.com>
Closes#11836 from andrewor14/use-session-catalog.
## What changes were proposed in this pull request?
1. Deprecated unionAll. It is pretty confusing to have both "union" and "unionAll" when the two do the same thing in Spark but are different in SQL.
2. Rename reduce in KeyValueGroupedDataset to reduceGroups so it is more consistent with rest of the functions in KeyValueGroupedDataset. Also makes it more obvious what "reduce" and "reduceGroups" mean. Previously it was confusing because it could be reducing a Dataset, or just reducing groups.
3. Added a "name" function, which is more natural to name columns than "as" for non-SQL users.
4. Remove "subtract" function since it is just an alias for "except".
## How was this patch tested?
All changes should be covered by existing tests. Also added couple test cases to cover "name".
Author: Reynold Xin <rxin@databricks.com>
Closes#11908 from rxin/SPARK-14088.
Building on the `SerializerManager` introduced in SPARK-13926/ #11755, this patch Spark modifies Spark's BlockManager to use RDD's ClassTags in order to select the best serializer to use when caching RDD blocks.
When storing a local block, the BlockManager `put()` methods use implicits to record ClassTags and stores those tags in the blocks' BlockInfo records. When reading a local block, the stored ClassTag is used to pick the appropriate serializer. When a block is stored with replication, the class tag is written into the block transfer metadata and will also be stored in the remote BlockManager.
There are two or three places where we don't properly pass ClassTags, including TorrentBroadcast and BlockRDD. I think this happens to work because the missing ClassTag always happens to be `ClassTag.Any`, but it might be worth looking more carefully at those places to see whether we should be more explicit.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#11801 from JoshRosen/pick-best-serializer-for-caching.
## What changes were proposed in this pull request?
This patch merges DatasetHolder and DataFrameHolder. This makes more sense because DataFrame/Dataset are now one class.
In addition, fixed some minor issues with pull request #11732.
## How was this patch tested?
Updated existing unit tests that test these implicits.
Author: Reynold Xin <rxin@databricks.com>
Closes#11737 from rxin/SPARK-13898.
## What changes were proposed in this pull request?
Previously, Dataset.groupBy returns a GroupedData, and Dataset.groupByKey returns a GroupedDataset. The naming is very similar, and unfortunately does not convey the real differences between the two.
Assume we are grouping by some keys (K). groupByKey is a key-value style group by, in which the schema of the returned dataset is a tuple of just two fields: key and value. groupBy, on the other hand, is a relational style group by, in which the schema of the returned dataset is flattened and contain |K| + |V| fields.
This pull request also removes the experimental tag from RelationalGroupedDataset. It has been with DataFrame since 1.3, and we have enough confidence now to stabilize it.
## How was this patch tested?
This is a rename to improve API understandability. Should be covered by all existing tests.
Author: Reynold Xin <rxin@databricks.com>
Closes#11841 from rxin/SPARK-13897.
## What changes were proposed in this pull request?
Logging was made private in Spark 2.0. If we move it, then users would be able to create a Logging trait themselves to avoid changing their own code.
## How was this patch tested?
existing tests.
Author: Wenchen Fan <wenchen@databricks.com>
Closes#11764 from cloud-fan/logger.
MiMa excludes are currently generated using both the current Spark version's classes and Spark 1.2.0's classes, but this doesn't make sense: we should only be ignoring classes which were `private` in the previous Spark version, not classes which became private in the current version.
This patch updates `dev/mima` to only generate excludes with respect to the previous artifacts that MiMa checks against. It also updates `MimaBuild` so that `excludeClass` only applies directly to the class being excluded and not to its companion object (since a class and its companion object can have different accessibility).
Author: Josh Rosen <joshrosen@databricks.com>
Closes#11774 from JoshRosen/SPARK-13948.
Because ClassTags are available when constructing ShuffledRDD we can use them to automatically use Kryo for shuffle serialization when the RDD's types are known to be compatible with Kryo.
This patch introduces `SerializerManager`, a component which picks the "best" serializer for a shuffle given the elements' ClassTags. It will automatically pick a Kryo serializer for ShuffledRDDs whose key, value, and/or combiner types are primitives, arrays of primitives, or strings. In the future we can use this class as a narrow extension point to integrate specialized serializers for other types, such as ByteBuffers.
In a planned followup patch, I will extend the BlockManager APIs so that we're able to use similar automatic serializer selection when caching RDDs (this is a little trickier because the ClassTags need to be threaded through many more places).
Author: Josh Rosen <joshrosen@databricks.com>
Closes#11755 from JoshRosen/automatically-pick-best-serializer.
## What changes were proposed in this pull request?
This PR removes three minor duplicated lines. First one is making the following unreachable code warning.
```
JoinSuite.scala:52: unreachable code
[warn] case j: BroadcastHashJoin => j
```
The other two are just consecutive repetitions in `Seq` of MiMa filters.
## How was this patch tested?
Pass the existing Jenkins test.
Author: Dongjoon Hyun <dongjoon@apache.org>
Closes#11773 from dongjoon-hyun/remove_duplicated_line.
## What changes were proposed in this pull request?
Narrow down the parameter type of `UserDefinedType#serialize()`. Currently, the parameter type is `Any`, however it would logically make more sense to narrow it down to the type of the actual user defined type.
## How was this patch tested?
Existing tests were successfully run on local machine.
Author: Jakob Odersky <jakob@odersky.com>
Closes#11379 from jodersky/SPARK-11011-udt-types.
## What changes were proposed in this pull request?
Add row/column iterator to local matrices to simplify tasks like BlockMatrix => RowMatrix conversion. It handles dense and sparse matrices properly.
## How was this patch tested?
Unit tests on sparse and dense matrix.
cc: dbtsai
Author: Xiangrui Meng <meng@databricks.com>
Closes#11757 from mengxr/SPARK-13927.
## What changes were proposed in this pull request?
We are able to change `Experimental` and `DeveloperAPI` API freely but also should monitor and manage those API carefully. This PR for [SPARK-13920](https://issues.apache.org/jira/browse/SPARK-13920) enables MiMa check and adds filters for them.
## How was this patch tested?
Pass the Jenkins tests (including MiMa).
Author: Dongjoon Hyun <dongjoon@apache.org>
Closes#11751 from dongjoon-hyun/SPARK-13920.
## What changes were proposed in this pull request?
1. Rename DataFrame.scala Dataset.scala, since the class is now named Dataset.
2. Remove LegacyFunctions. It was introduced in Spark 1.6 for backward compatibility, and can be removed in Spark 2.0.
## How was this patch tested?
Should be covered by existing unit/integration tests.
Author: Reynold Xin <rxin@databricks.com>
Closes#11704 from rxin/SPARK-13880.
## What changes were proposed in this pull request?
`LinearRegressionWithSGD` and `StreamingLinearRegressionWithSGD` does not have `regParam` as their constructor arguments. They just depends on GradientDescent's default reqParam values.
To be consistent with other algorithms, we had better add them. The same default value is used.
## How was this patch tested?
Pass the existing unit test.
Author: Dongjoon Hyun <dongjoon@apache.org>
Closes#11527 from dongjoon-hyun/SPARK-13686.
## What changes were proposed in this pull request?
This PR fixes 135 typos over 107 files:
* 121 typos in comments
* 11 typos in testcase name
* 3 typos in log messages
## How was this patch tested?
Manual.
Author: Dongjoon Hyun <dongjoon@apache.org>
Closes#11689 from dongjoon-hyun/fix_more_typos.
## What changes were proposed in this pull request?
For 2.0.0, we had better make **sbt** and **sbt plugins** up-to-date. This PR checks the status of each plugins and bumps the followings.
* sbt: 0.13.9 --> 0.13.11
* sbteclipse-plugin: 2.2.0 --> 4.0.0
* sbt-dependency-graph: 0.7.4 --> 0.8.2
* sbt-mima-plugin: 0.1.6 --> 0.1.9
* sbt-revolver: 0.7.2 --> 0.8.0
All other plugins are up-to-date. (Note that `sbt-avro` seems to be change from 0.3.2 to 1.0.1, but it's not published in the repository.)
During upgrade, this PR also updated the following MiMa error. Note that the related excluding filter is already registered correctly. It seems due to the change of MiMa exception result.
```
// SPARK-12896 Send only accumulator updates to driver, not TaskMetrics
ProblemFilters.exclude[IncompatibleMethTypeProblem]("org.apache.spark.Accumulable.this"),
-ProblemFilters.exclude[IncompatibleMethTypeProblem]("org.apache.spark.Accumulator.this"),
+ProblemFilters.exclude[DirectMissingMethodProblem]("org.apache.spark.Accumulator.this"),
```
## How was this patch tested?
Pass the Jenkins build.
Author: Dongjoon Hyun <dongjoon@apache.org>
Closes#11669 from dongjoon-hyun/update_mima.
## What changes were proposed in this pull request?
PR #11443 temporarily disabled MiMA check, this PR re-enables it.
One extra change is that `object DataFrame` is also removed. The only purpose of introducing `object DataFrame` was to use it as an internal factory for creating `Dataset[Row]`. By replacing this internal factory with `Dataset.newDataFrame`, both `DataFrame` and `DataFrame$` are entirely removed from the API, so that we can simply put a `MissingClassProblem` filter in `MimaExcludes.scala` for most DataFrame API changes.
## How was this patch tested?
Tested by MiMA check triggered by Jenkins.
Author: Cheng Lian <lian@databricks.com>
Closes#11656 from liancheng/re-enable-mima.
`HadoopFsRelation` is used for reading most files into Spark SQL. However today this class mixes the concerns of file management, schema reconciliation, scan building, bucketing, partitioning, and writing data. As a result, many data sources are forced to reimplement the same functionality and the various layers have accumulated a fair bit of inefficiency. This PR is a first cut at separating this into several components / interfaces that are each described below. Additionally, all implementations inside of Spark (parquet, csv, json, text, orc, svmlib) have been ported to the new API `FileFormat`. External libraries, such as spark-avro will also need to be ported to work with Spark 2.0.
### HadoopFsRelation
A simple `case class` that acts as a container for all of the metadata required to read from a datasource. All discovery, resolution and merging logic for schemas and partitions has been removed. This an internal representation that no longer needs to be exposed to developers.
```scala
case class HadoopFsRelation(
sqlContext: SQLContext,
location: FileCatalog,
partitionSchema: StructType,
dataSchema: StructType,
bucketSpec: Option[BucketSpec],
fileFormat: FileFormat,
options: Map[String, String]) extends BaseRelation
```
### FileFormat
The primary interface that will be implemented by each different format including external libraries. Implementors are responsible for reading a given format and converting it into `InternalRow` as well as writing out an `InternalRow`. A format can optionally return a schema that is inferred from a set of files.
```scala
trait FileFormat {
def inferSchema(
sqlContext: SQLContext,
options: Map[String, String],
files: Seq[FileStatus]): Option[StructType]
def prepareWrite(
sqlContext: SQLContext,
job: Job,
options: Map[String, String],
dataSchema: StructType): OutputWriterFactory
def buildInternalScan(
sqlContext: SQLContext,
dataSchema: StructType,
requiredColumns: Array[String],
filters: Array[Filter],
bucketSet: Option[BitSet],
inputFiles: Array[FileStatus],
broadcastedConf: Broadcast[SerializableConfiguration],
options: Map[String, String]): RDD[InternalRow]
}
```
The current interface is based on what was required to get all the tests passing again, but still mixes a couple of concerns (i.e. `bucketSet` is passed down to the scan instead of being resolved by the planner). Additionally, scans are still returning `RDD`s instead of iterators for single files. In a future PR, bucketing should be removed from this interface and the scan should be isolated to a single file.
### FileCatalog
This interface is used to list the files that make up a given relation, as well as handle directory based partitioning.
```scala
trait FileCatalog {
def paths: Seq[Path]
def partitionSpec(schema: Option[StructType]): PartitionSpec
def allFiles(): Seq[FileStatus]
def getStatus(path: Path): Array[FileStatus]
def refresh(): Unit
}
```
Currently there are two implementations:
- `HDFSFileCatalog` - based on code from the old `HadoopFsRelation`. Infers partitioning by recursive listing and caches this data for performance
- `HiveFileCatalog` - based on the above, but it uses the partition spec from the Hive Metastore.
### ResolvedDataSource
Produces a logical plan given the following description of a Data Source (which can come from DataFrameReader or a metastore):
- `paths: Seq[String] = Nil`
- `userSpecifiedSchema: Option[StructType] = None`
- `partitionColumns: Array[String] = Array.empty`
- `bucketSpec: Option[BucketSpec] = None`
- `provider: String`
- `options: Map[String, String]`
This class is responsible for deciding which of the Data Source APIs a given provider is using (including the non-file based ones). All reconciliation of partitions, buckets, schema from metastores or inference is done here.
### DataSourceAnalysis / DataSourceStrategy
Responsible for analyzing and planning reading/writing of data using any of the Data Source APIs, including:
- pruning the files from partitions that will be read based on filters.
- appending partition columns*
- applying additional filters when a data source can not evaluate them internally.
- constructing an RDD that is bucketed correctly when required*
- sanity checking schema match-up and other analysis when writing.
*In the future we should do that following:
- Break out file handling into its own Strategy as its sufficiently complex / isolated.
- Push the appending of partition columns down in to `FileFormat` to avoid an extra copy / unvectorization.
- Use a custom RDD for scans instead of `SQLNewNewHadoopRDD2`
Author: Michael Armbrust <michael@databricks.com>
Author: Wenchen Fan <wenchen@databricks.com>
Closes#11509 from marmbrus/fileDataSource.
…ing partitions
I'm pretty sure this is the reason we couldn't easily recover from an unbalanced Kafka partition under heavy load when using backpressure.
`maxMessagesPerPartition` calculates an appropriate limit for the message rate from all partitions, and then divides by the number of partitions to determine how many messages to retrieve per partition. The problem with this approach is that when one partition is behind by millions of records (due to random Kafka issues), but the rate estimator calculates only 100k total messages can be retrieved, each partition (out of say 32) only retrieves max 100k/32=3125 messages.
This PR (still needing a test) determines a per-partition desired message count by using the current lag for each partition to preferentially weight the total message limit among the partitions. In this situation, if each partition gets 1k messages, but 1 partition starts 1M behind, then the total number of messages to retrieve is (32 * 1k + 1M) = 1032000 messages, of which the one partition needs 1001000. So, it gets (1001000 / 1032000) = 97% of the 100k messages, and the other 31 partitions share the remaining 3%.
Assuming all of 100k the messages are retrieved and processed within the batch window, the rate calculator will increase the number of messages to retrieve in the next batch, until it reaches a new stable point or the backlog is finished processed.
We're going to try deploying this internally at Shopify to see if this resolves our issue.
tdas koeninger holdenk
Author: Jason White <jason.white@shopify.com>
Closes#10089 from JasonMWhite/rate_controller_offsets.
## What changes were proposed in this pull request?
This creates a `SessionState`, which groups a few fields that existed in `SQLContext`. Because `HiveContext` extends `SQLContext` we also need to make changes there. This is mainly a cleanup task that will soon pave the way for merging the two contexts.
## How was this patch tested?
Existing unit tests; this patch introduces no change in behavior.
Author: Andrew Or <andrew@databricks.com>
Closes#11405 from andrewor14/refactor-session.
## What changes were proposed in this pull request?
TaskContext supports task completion callback, which gets called regardless of task failures. However, there is no way for the listener to know if there is an error. This patch adds a new listener that gets called when a task fails.
## How was the this patch tested?
New unit test case and integration test case covering the code path
Author: Reynold Xin <rxin@databricks.com>
Closes#11340 from rxin/SPARK-13465.
## What changes were proposed in this pull request?
This patch moves SQLConf into org.apache.spark.sql.internal package to make it very explicit that it is internal. Soon I will also submit more API work that creates implementations of interfaces in this internal package.
## How was this patch tested?
If it compiles, then the refactoring should work.
Author: Reynold Xin <rxin@databricks.com>
Closes#11363 from rxin/SPARK-13486.
## What changes were proposed in this pull request?
This patch removes SparkContext.metricsSystem. SparkContext.metricsSystem returns MetricsSystem, which is a private class. I think it was added by accident.
In addition, I also removed an unused private[spark] method schedulerBackend setter.
## How was the this patch tested?
N/A.
Author: Reynold Xin <rxin@databricks.com>
This patch had conflicts when merged, resolved by
Committer: Josh Rosen <joshrosen@databricks.com>
Closes#11282 from rxin/SPARK-13413.
## What changes were proposed in this pull request?
This PR removes the support of SIMR, since SIMR is not actively used and maintained for a long time, also is not supported from `SparkSubmit`, so here propose to remove it.
## How was the this patch tested?
This patch is tested locally by running unit tests.
Author: jerryshao <sshao@hortonworks.com>
Closes#11296 from jerryshao/SPARK-13426.
This pull request has the following changes:
1. Moved UserDefinedFunction into expressions package. This is more consistent with how we structure the packages for window functions and UDAFs.
2. Moved UserDefinedPythonFunction into execution.python package, so we don't have a random private class in the top level sql package.
3. Move everything in execution/python.scala into the newly created execution.python package.
Most of the diffs are just straight copy-paste.
Author: Reynold Xin <rxin@databricks.com>
Closes#11181 from rxin/SPARK-13296.
When the HistoryServer is showing an incomplete app, it needs to check if there is a newer version of the app available. It does this by checking if a version of the app has been loaded with a larger *filesize*. If so, it detaches the current UI, attaches the new one, and redirects back to the same URL to show the new UI.
https://issues.apache.org/jira/browse/SPARK-7889
Author: Steve Loughran <stevel@hortonworks.com>
Author: Imran Rashid <irashid@cloudera.com>
Closes#11118 from squito/SPARK-7889-alternate.
This PR improve the lookup of BytesToBytesMap by:
1. Generate code for calculate the hash code of grouping keys.
2. Do not use MemoryLocation, fetch the baseObject and offset for key and value directly (remove the indirection).
Author: Davies Liu <davies@databricks.com>
Closes#11010 from davies/gen_map.
Additional changes to #10835, mainly related to style and visibility. This patch also adds back a few deprecated methods for backward compatibility.
Author: Andrew Or <andrew@databricks.com>
Closes#10958 from andrewor14/task-metrics-to-accums-followups.
JIRA: https://issues.apache.org/jira/browse/SPARK-12689
DDLParser processes three commands: createTable, describeTable and refreshTable.
This patch migrates the three commands to newly absorbed parser.
Author: Liang-Chi Hsieh <viirya@gmail.com>
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes#10723 from viirya/migrate-ddl-describe.
This patch changes Spark's build to make Scala 2.11 the default Scala version. To be clear, this does not mean that Spark will stop supporting Scala 2.10: users will still be able to compile Spark for Scala 2.10 by following the instructions on the "Building Spark" page; however, it does mean that Scala 2.11 will be the default Scala version used by our CI builds (including pull request builds).
The Scala 2.11 compiler is faster than 2.10, so I think we'll be able to look forward to a slight speedup in our CI builds (it looks like it's about 2X faster for the Maven compile-only builds, for instance).
After this patch is merged, I'll update Jenkins to add new compile-only jobs to ensure that Scala 2.10 compilation doesn't break.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#10608 from JoshRosen/SPARK-6363.
[SPARK-10873] Support column sort and search for History Server using jQuery DataTable and REST API. Before this commit, the history server was generated hard-coded html and can not support search, also, the sorting was disabled if there is any application that has more than one attempt. Supporting search and sort (over all applications rather than the 20 entries in the current page) in any case will greatly improve user experience.
1. Create the historypage-template.html for displaying application information in datables.
2. historypage.js uses jQuery to access the data from /api/v1/applications REST API, and use DataTable to display each application's information. For application that has more than one attempt, the RowsGroup is used to merge such entries while at the same time supporting sort and search.
3. "duration" and "lastUpdated" rest API are added to application's "attempts".
4. External javascirpt and css files for datatables, RowsGroup and jquery plugins are added with licenses clarified.
Snapshots for how it looks like now:
History page view:
Search:
Sort by started time:
Author: zhuol <zhuol@yahoo-inc.com>
Closes#10648 from zhuoliu/10873.
The high level idea is that instead of having the executors send both accumulator updates and TaskMetrics, we should have them send only accumulator updates. This eliminates the need to maintain both code paths since one can be implemented in terms of the other. This effort is split into two parts:
**SPARK-12895: Implement TaskMetrics using accumulators.** TaskMetrics is basically just a bunch of accumulable fields. This patch makes TaskMetrics a syntactic wrapper around a collection of accumulators so we don't need to send TaskMetrics from the executors to the driver.
**SPARK-12896: Send only accumulator updates to the driver.** Now that TaskMetrics are expressed in terms of accumulators, we can capture all TaskMetrics values if we just send accumulator updates from the executors to the driver. This completes the parent issue SPARK-10620.
While an effort has been made to preserve as much of the public API as possible, there were a few known breaking DeveloperApi changes that would be very awkward to maintain. I will gather the full list shortly and post it here.
Note: This was once part of #10717. This patch is split out into its own patch from there to make it easier for others to review. Other smaller pieces of already been merged into master.
Author: Andrew Or <andrew@databricks.com>
Closes#10835 from andrewor14/task-metrics-use-accums.
… Add LibSVMOutputWriter
The behavior of LibSVMRelation is not changed except adding LibSVMOutputWriter
* Partition is still not supported
* Multiple input paths is not supported
Author: Jeff Zhang <zjffdu@apache.org>
Closes#9595 from zjffdu/SPARK-11622.
Fix Java function API methods for flatMap and mapPartitions to require producing only an Iterator, not Iterable. Also fix DStream.flatMap to require a function producing TraversableOnce only, not Traversable.
CC rxin pwendell for API change; tdas since it also touches streaming.
Author: Sean Owen <sowen@cloudera.com>
Closes#10413 from srowen/SPARK-3369.
Added color coding to the Executors page for Active Tasks, Failed Tasks, Completed Tasks and Task Time.
Active Tasks is shaded blue with it's range based on percentage of total cores used.
Failed Tasks is shaded red ranging over the first 10% of total tasks failed
Completed Tasks is shaded green ranging over 10% of total tasks including failed and active tasks, but only when there are active or failed tasks on that executor.
Task Time is shaded red when GC Time goes over 10% of total time with it's range directly corresponding to the percent of total time.
Author: Alex Bozarth <ajbozart@us.ibm.com>
Closes#10154 from ajbozarth/spark12149.
- Remove Akka dependency from core. Note: the streaming-akka project still uses Akka.
- Remove HttpFileServer
- Remove Akka configs from SparkConf and SSLOptions
- Rename `spark.akka.frameSize` to `spark.rpc.message.maxSize`. I think it's still worth to keep this config because using `DirectTaskResult` or `IndirectTaskResult` depends on it.
- Update comments and docs
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10854 from zsxwing/remove-akka.
Include the following changes:
1. Add "streaming-akka" project and org.apache.spark.streaming.akka.AkkaUtils for creating an actorStream
2. Remove "StreamingContext.actorStream" and "JavaStreamingContext.actorStream"
3. Update the ActorWordCount example and add the JavaActorWordCount example
4. Make "streaming-zeromq" depend on "streaming-akka" and update the codes accordingly
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10744 from zsxwing/streaming-akka-2.
Including the following changes:
1. Add StreamingListenerForwardingBus to WrappedStreamingListenerEvent process events in `onOtherEvent` to StreamingListener
2. Remove StreamingListenerBus
3. Merge AsynchronousListenerBus and LiveListenerBus to the same class LiveListenerBus
4. Add `logEvent` method to SparkListenerEvent so that EventLoggingListener can use it to ignore WrappedStreamingListenerEvents
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10779 from zsxwing/streaming-listener.
This pull request removes the public developer parser API for external parsers. Given everything a parser depends on (e.g. logical plans and expressions) are internal and not stable, external parsers will break with every release of Spark. It is a bad idea to create the illusion that Spark actually supports pluggable parsers. In addition, this also reduces incentives for 3rd party projects to contribute parse improvements back to Spark.
Author: Reynold Xin <rxin@databricks.com>
Closes#10801 from rxin/SPARK-12855.
This pull request removes the external block store API. This is rarely used, and the file system interface is actually a better, more standard way to interact with external storage systems.
There are some other things to remove also, as pointed out by JoshRosen. We will do those as follow-up pull requests.
Author: Reynold Xin <rxin@databricks.com>
Closes#10752 from rxin/remove-offheap.
Fix the style violation (space before , and :).
This PR is a followup for #10643.
Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>
Closes#10685 from sarutak/SPARK-12692-followup-streaming.
Replace Guava `Optional` with (an API clone of) Java 8 `java.util.Optional` (edit: and a clone of Guava `Optional`)
See also https://github.com/apache/spark/pull/10512
Author: Sean Owen <sowen@cloudera.com>
Closes#10513 from srowen/SPARK-4819.
The default serializer in Kryo is FieldSerializer and it ignores transient fields and never calls `writeObject` or `readObject`. So we should register OpenHashMapBasedStateMap using `DefaultSerializer` to make it work with Kryo.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10609 from zsxwing/SPARK-12591.
This PR includes the following changes:
1. Rename `ActorReceiver` to `ActorReceiverSupervisor`
2. Remove `ActorHelper`
3. Add a new `ActorReceiver` for Scala and `JavaActorReceiver` for Java
4. Add `JavaActorWordCount` example
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10457 from zsxwing/java-actor-stream.
Whole code of Vector.scala, VectorSuite.scala and GraphKryoRegistrator.scala are no longer used so it's time to remove them in Spark 2.0.
Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>
Closes#10613 from sarutak/SPARK-12665.
Cartesian product use UnsafeExternalSorter without comparator to do spilling, it will NPE if spilling happens.
This bug also hitted by #10605
cc JoshRosen
Author: Davies Liu <davies@databricks.com>
Closes#10606 from davies/fix_spilling.
I looked at each case individually and it looks like they can all be removed. The only one that I had to think twice was toArray (I even thought about un-deprecating it, until I realized it was a problem in Java to have toArray returning java.util.List).
Author: Reynold Xin <rxin@databricks.com>
Closes#10569 from rxin/SPARK-12615.
### Remove AkkaRpcEnv
Keep `SparkEnv.actorSystem` because Streaming still uses it. Will remove it and AkkaUtils after refactoring Streaming actorStream API.
### Remove systemName
There are 2 places using `systemName`:
* `RpcEnvConfig.name`. Actually, although it's used as `systemName` in `AkkaRpcEnv`, `NettyRpcEnv` uses it as the service name to output the log `Successfully started service *** on port ***`. Since the service name in log is useful, I keep `RpcEnvConfig.name`.
* `def setupEndpointRef(systemName: String, address: RpcAddress, endpointName: String)`. Each `ActorSystem` has a `systemName`. Akka requires `systemName` in its URI and will refuse a connection if `systemName` is not matched. However, `NettyRpcEnv` doesn't use it. So we can remove `systemName` from `setupEndpointRef` since we are removing `AkkaRpcEnv`.
### Remove RpcEnv.uriOf
`uriOf` exists because Akka uses different URI formats for with and without authentication, e.g., `akka.ssl.tcp...` and `akka.tcp://...`. But `NettyRpcEnv` uses the same format. So it's not necessary after removing `AkkaRpcEnv`.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#10459 from zsxwing/remove-akka-rpc-env.
We switched to TorrentBroadcast in Spark 1.1, and HttpBroadcast has been undocumented since then. It's time to remove it in Spark 2.0.
Author: Reynold Xin <rxin@databricks.com>
Closes#10531 from rxin/SPARK-12588.
Add `computePrincipalComponentsAndVariance` to also compute PCA's explained variance.
CC mengxr
Author: Sean Owen <sowen@cloudera.com>
Closes#9736 from srowen/SPARK-11530.
The json endpoint for stages doesn't include information on the stage duration that is present in the UI. This looks like a simple oversight, they should be included. eg., the metrics should be included at api/v1/applications/<appId>/stages.
Metrics I've added are: submissionTime, firstTaskLaunchedTime and completionTime
Author: Xin Ren <iamshrek@126.com>
Closes#10107 from keypointt/SPARK-11155.
I have tried to address all the comments in pull request https://github.com/apache/spark/pull/2447.
Note that the second commit (using the new method in all internal code of all components) is quite intrusive and could be omitted.
Author: Jeroen Schot <jeroen.schot@surfsara.nl>
Closes#9767 from schot/master.
In the previous implementation, the driver needs to know the executor listening address to send the thread dump request. However, in Netty RPC, the executor doesn't listen to any port, so the executor thread dump feature is broken.
This patch makes the driver use the endpointRef stored in BlockManagerMasterEndpoint to send the thread dump request to fix it.
Author: Shixiong Zhu <shixiong@databricks.com>
Closes#9976 from zsxwing/executor-thread-dump.
Currently streaming foreachRDD Java API uses a function prototype requiring a return value of null. This PR deprecates the old method and uses VoidFunction to allow for more concise declaration. Also added VoidFunction2 to Java API in order to use in Streaming methods. Unit test is added for using foreachRDD with VoidFunction, and changes have been tested with Java 7 and Java 8 using lambdas.
Author: Bryan Cutler <bjcutler@us.ibm.com>
Closes#9488 from BryanCutler/foreachRDD-VoidFunction-SPARK-4557.
This adds an extra filter for private or protected classes. We only filter for package private right now.
Author: Timothy Hunter <timhunter@databricks.com>
Closes#9697 from thunterdb/spark-11732.
This is to support JSON serialization of Param[Vector] in the pipeline API. It could be used for other purposes too. The schema is the same as `VectorUDT`. jkbradley
Author: Xiangrui Meng <meng@databricks.com>
Closes#9751 from mengxr/SPARK-11766.
Like ml ```LinearRegression```, ```LogisticRegression``` should provide a training summary including feature names and their coefficients.
Author: Yanbo Liang <ybliang8@gmail.com>
Closes#9303 from yanboliang/spark-9492.
Since we do not need to preserve a page before calling compute(), MapPartitionsWithPreparationRDD is not needed anymore.
This PR basically revert #8543, #8511, #8038, #8011
Author: Davies Liu <davies@databricks.com>
Closes#9381 from davies/remove_prepare2.
There's a lot of duplication between SortShuffleManager and UnsafeShuffleManager. Given that these now provide the same set of functionality, now that UnsafeShuffleManager supports large records, I think that we should replace SortShuffleManager's serialized shuffle implementation with UnsafeShuffleManager's and should merge the two managers together.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#8829 from JoshRosen/consolidate-sort-shuffle-implementations.
This PR improve the sessions management by replacing the thread-local based to one SQLContext per session approach, introduce separated temporary tables and UDFs/UDAFs for each session.
A new session of SQLContext could be created by:
1) create an new SQLContext
2) call newSession() on existing SQLContext
For HiveContext, in order to reduce the cost for each session, the classloader and Hive client are shared across multiple sessions (created by newSession).
CacheManager is also shared by multiple sessions, so cache a table multiple times in different sessions will not cause multiple copies of in-memory cache.
Added jars are still shared by all the sessions, because SparkContext does not support sessions.
cc marmbrus yhuai rxin
Author: Davies Liu <davies@databricks.com>
Closes#8909 from davies/sessions.
This PR remove the typeId in columnar cache, it's not needed anymore, it also remove DATE and TIMESTAMP (use INT/LONG instead).
Author: Davies Liu <davies@databricks.com>
Closes#8989 from davies/refactor_cache.
In many modeling application, data points are not necessarily sampled with equal probabilities. Linear regression should support weighting which account the over or under sampling.
work in progress.
Author: Meihua Wu <meihuawu@umich.edu>
Closes#8631 from rotationsymmetry/SPARK-9642.
When speculative execution is enabled, consider a scenario where the authorized committer of a particular output partition fails during the OutputCommitter.commitTask() call. In this case, the OutputCommitCoordinator is supposed to release that committer's exclusive lock on committing once that task fails. However, due to a unit mismatch (we used task attempt number in one place and task attempt id in another) the lock will not be released, causing Spark to go into an infinite retry loop.
This bug was masked by the fact that the OutputCommitCoordinator does not have enough end-to-end tests (the current tests use many mocks). Other factors contributing to this bug are the fact that we have many similarly-named identifiers that have different semantics but the same data types (e.g. attemptNumber and taskAttemptId, with inconsistent variable naming which makes them difficult to distinguish).
This patch adds a regression test and fixes this bug by always using task attempt numbers throughout this code.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#8544 from JoshRosen/SPARK-10381.
In fraud detection dataset, almost all the samples are negative while only couple of them are positive. This type of high imbalanced data will bias the models toward negative resulting poor performance. In python-scikit, they provide a correction allowing users to Over-/undersample the samples of each class according to the given weights. In auto mode, selects weights inversely proportional to class frequencies in the training set. This can be done in a more efficient way by multiplying the weights into loss and gradient instead of doing actual over/undersampling in the training dataset which is very expensive.
http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html
On the other hand, some of the training data maybe more important like the training samples from tenure users while the training samples from new users maybe less important. We should be able to provide another "weight: Double" information in the LabeledPoint to weight them differently in the learning algorithm.
Author: DB Tsai <dbt@netflix.com>
Author: DB Tsai <dbt@dbs-mac-pro.corp.netflix.com>
Closes#7884 from dbtsai/SPARK-7685.
We introduced the Netty network module for shuffle in Spark 1.2, and has turned it on by default for 3 releases. The old ConnectionManager is difficult to maintain. If we merge the patch now, by the time it is released, it would be 1 yr for which ConnectionManager is off by default. It's time to remove it.
Author: Reynold Xin <rxin@databricks.com>
Closes#8161 from rxin/SPARK-9767.
To correctly isolate applications, when requests to read shuffle data
arrive at the shuffle service, proper authorization checks need to
be performed. This change makes sure that only the application that
created the shuffle data can read from it.
Such checks are only enabled when "spark.authenticate" is enabled,
otherwise there's no secure way to make sure that the client is really
who it says it is.
Author: Marcelo Vanzin <vanzin@cloudera.com>
Closes#8218 from vanzin/SPARK-10004.
A fundamental limitation of the existing SQL tests is that *there is simply no way to create your own `SparkContext`*. This is a serious limitation because the user may wish to use a different master or config. As a case in point, `BroadcastJoinSuite` is entirely commented out because there is no way to make it pass with the existing infrastructure.
This patch removes the singletons `TestSQLContext` and `TestData`, and instead introduces a `SharedSQLContext` that starts a context per suite. Unfortunately the singletons were so ingrained in the SQL tests that this patch necessarily needed to touch *all* the SQL test files.
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/apache/spark/8111)
<!-- Reviewable:end -->
Author: Andrew Or <andrew@databricks.com>
Closes#8111 from andrewor14/sql-tests-refactor.
Made ProbabilisticClassifier, Identifiable, VectorUDT public. All are annotated as DeveloperApi.
CC: mengxr EronWright
Author: Joseph K. Bradley <joseph@databricks.com>
Closes#8004 from jkbradley/ml-api-public-items and squashes the following commits:
7ebefda [Joseph K. Bradley] update per code review
7ff0768 [Joseph K. Bradley] attepting to add mima fix
756d84c [Joseph K. Bradley] VectorUDT annotated as AlphaComponent
ae7767d [Joseph K. Bradley] added another warning
94fd553 [Joseph K. Bradley] Made ProbabilisticClassifier, Identifiable, VectorUDT public APIs
There are a few changes in this pull request:
1. Moved all data sources to execution.datasources, except the public JDBC APIs.
2. In order to maintain backward compatibility from 1, added a backward compatibility translation map in data source resolution.
3. Moved ui and metric package into execution.
4. Added more documentation on some internal classes.
5. Renamed DataSourceRegister.format -> shortName.
6. Added "override" modifier on shortName.
7. Removed IntSQLMetric.
Author: Reynold Xin <rxin@databricks.com>
Closes#8056 from rxin/SPARK-9763 and squashes the following commits:
9df4801 [Reynold Xin] Removed hardcoded name in test cases.
d9babc6 [Reynold Xin] Shorten.
e484419 [Reynold Xin] Removed VisibleForTesting.
171b812 [Reynold Xin] MimaExcludes.
2041389 [Reynold Xin] Compile ...
79dda42 [Reynold Xin] Compile.
0818ba3 [Reynold Xin] Removed IntSQLMetric.
c46884f [Reynold Xin] Two more fixes.
f9aa88d [Reynold Xin] [SPARK-9763][SQL] Minimize exposure of internal SQL classes.
https://issues.apache.org/jira/browse/SPARK-9602
Although we have hidden Akka behind RPC interface, I found that the Akka/Actor-related comments are still spreading everywhere. To make it consistent, we shall remove "actor"/"akka" words from the comments...
Author: CodingCat <zhunansjtu@gmail.com>
Closes#7936 from CodingCat/SPARK-9602 and squashes the following commits:
e8296a3 [CodingCat] remove actor words from comments
Certain use cases of Spark involve RDDs with long lineages that must be truncated periodically (e.g. GraphX). The existing way of doing it is through `rdd.checkpoint()`, which is expensive because it writes to HDFS. This patch provides an alternative to truncate lineages cheaply *without providing the same level of fault tolerance*.
**Local checkpointing** writes checkpointed data to the local file system through the block manager. It is much faster than replicating to a reliable storage and provides the same semantics as long as executors do not fail. It is accessible through a new operator `rdd.localCheckpoint()` and leaves the old one unchanged. Users may even decide to combine the two and call the reliable one less frequently.
The bulk of this patch involves refactoring the checkpointing interface to accept custom implementations of checkpointing. [Design doc](https://issues.apache.org/jira/secure/attachment/12741708/SPARK-7292-design.pdf).
Author: Andrew Or <andrew@databricks.com>
Closes#7279 from andrewor14/local-checkpoint and squashes the following commits:
729600f [Andrew Or] Oops, fix tests
34bc059 [Andrew Or] Avoid computing all partitions in local checkpoint
e43bbb6 [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
3be5aea [Andrew Or] Address comments
bf846a6 [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
ab003a3 [Andrew Or] Fix compile
c2e111b [Andrew Or] Address comments
33f167a [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
e908a42 [Andrew Or] Fix tests
f5be0f3 [Andrew Or] Use MEMORY_AND_DISK as the default local checkpoint level
a92657d [Andrew Or] Update a few comments
e58e3e3 [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
4eb6eb1 [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
1bbe154 [Andrew Or] Simplify LocalCheckpointRDD
48a9996 [Andrew Or] Avoid traversing dependency tree + rewrite tests
62aba3f [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
db70dc2 [Andrew Or] Express local checkpointing through caching the original RDD
87d43c6 [Andrew Or] Merge branch 'master' of github.com:apache/spark into local-checkpoint
c449b38 [Andrew Or] Fix style
4a182f3 [Andrew Or] Add fine-grained tests for local checkpointing
53b363b [Andrew Or] Rename a few more awkwardly named methods (minor)
e4cf071 [Andrew Or] Simplify LocalCheckpointRDD + docs + clean ups
4880deb [Andrew Or] Fix style
d096c67 [Andrew Or] Fix mima
172cb66 [Andrew Or] Fix mima?
e53d964 [Andrew Or] Fix style
56831c5 [Andrew Or] Add a few warnings and clear exception messages
2e59646 [Andrew Or] Add local checkpoint clean up tests
4dbbab1 [Andrew Or] Refactor CheckpointSuite to test local checkpointing
4514dc9 [Andrew Or] Clean local checkpoint files through RDD cleanups
0477eec [Andrew Or] Rename a few methods with awkward names (minor)
2e902e5 [Andrew Or] First implementation of local checkpointing
8447454 [Andrew Or] Fix tests
4ac1896 [Andrew Or] Refactor checkpoint interface for modularity
Dynamic allocation is a feature that allows a Spark application to scale the number of executors up and down dynamically based on the workload. Support was first introduced in YARN since 1.2, and then extended to Mesos coarse-grained mode recently. Today, it is finally supported in standalone mode as well!
I tested this locally and it works as expected. This is WIP because unit tests are coming.
Author: Andrew Or <andrew@databricks.com>
Closes#7532 from andrewor14/standalone-da and squashes the following commits:
b3c1736 [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
879e928 [Andrew Or] Add end-to-end tests for standalone dynamic allocation
accc8f6 [Andrew Or] Address comments
ee686a8 [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
c0a2c02 [Andrew Or] Fix build after merge conflict
24149eb [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
2e762d6 [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
6832bd7 [Andrew Or] Add tests for scheduling with executor limit
a82e907 [Andrew Or] Fix comments
0a8be79 [Andrew Or] Simplify logic by removing the worker blacklist
b7742af [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
2eb5f3f [Andrew Or] Merge branch 'master' of github.com:apache/spark into standalone-da
1334e9a [Andrew Or] Fix MiMa
32abe44 [Andrew Or] Fix style
58cb06f [Andrew Or] Privatize worker blacklist for cleanliness
42ac215 [Andrew Or] Clean up comments and rewrite code for readability
49702d1 [Andrew Or] Clean up shuffle files after application exits
80047aa [Andrew Or] First working implementation
This way, the sources package contains only public facing interfaces.
Author: Reynold Xin <rxin@databricks.com>
Closes#7565 from rxin/move-ds and squashes the following commits:
7661aff [Reynold Xin] Mima
9d5196a [Reynold Xin] Rearranged imports.
3dd7174 [Reynold Xin] [SPARK-8906][SQL] Move all internal data source classes into execution.datasources.
This PR also remove the duplicated code between registerFunction and UserDefinedFunction.
cc JoshRosen
Author: Davies Liu <davies@databricks.com>
Closes#7450 from davies/fix_return_type and squashes the following commits:
e80bf9f [Davies Liu] remove debugging code
f94b1f6 [Davies Liu] fix mima
8f9c58b [Davies Liu] convert returned object from UDF into internal type
Modifying Vector, DenseVector, and SparseVector to implement argmax functionality. This work is to set the stage for changes to be done in Spark-7423.
Author: George Dittmar <georgedittmar@gmail.com>
Author: George <dittmar@Georges-MacBook-Pro.local>
Author: dittmarg <george.dittmar@webtrends.com>
Author: Xiangrui Meng <meng@databricks.com>
Closes#6112 from GeorgeDittmar/SPARK-7422 and squashes the following commits:
3e0a939 [George Dittmar] Merge pull request #1 from mengxr/SPARK-7422
127dec5 [Xiangrui Meng] update argmax impl
2ea6a55 [George Dittmar] Added MimaExcludes for Vectors.argmax
98058f4 [George Dittmar] Merge branch 'master' of github.com:apache/spark into SPARK-7422
5fd9380 [George Dittmar] fixing style check error
42341fb [George Dittmar] refactoring arg max check to better handle zero values
b22af46 [George Dittmar] Fixing spaces between commas in unit test
f2eba2f [George Dittmar] Cleaning up unit tests to be fewer lines
aa330e3 [George Dittmar] Fixing some last if else spacing issues
ac53c55 [George Dittmar] changing dense vector argmax unit test to be one line call vs 2
d5b5423 [George Dittmar] Fixing code style and updating if logic on when to check for zero values
ee1a85a [George Dittmar] Cleaning up unit tests a bit and modifying a few cases
3ee8711 [George Dittmar] Fixing corner case issue with zeros in the active values of the sparse vector. Updated unit tests
b1f059f [George Dittmar] Added comment before we start arg max calculation. Updated unit tests to cover corner cases
f21dcce [George Dittmar] commit
af17981 [dittmarg] Initial work fixing bug that was made clear in pr
eeda560 [George] Fixing SparseVector argmax function to ignore zero values while doing the calculation.
4526acc [George] Merge branch 'master' of github.com:apache/spark into SPARK-7422
df9538a [George] Added argmax to sparse vector and added unit test
3cffed4 [George] Adding unit tests for argmax functions for Dense and Sparse vectors
04677af [George] initial work on adding argmax to Vector and SparseVector
This PR enables SparkR to dynamically ship the SparkR binary package to the AM node in YARN cluster mode, thus it is no longer required that the SparkR package be installed on each worker node.
This PR uses the JDK jar tool to package the SparkR package, because jar is thought to be available on both Linux/Windows platforms where JDK has been installed.
This PR does not address the R worker involved in RDD API. Will address it in a separate JIRA issue.
This PR does not address SBT build. SparkR installation and packaging by SBT will be addressed in a separate JIRA issue.
R/install-dev.bat is not tested. shivaram , Could you help to test it?
Author: Sun Rui <rui.sun@intel.com>
Closes#6743 from sun-rui/SPARK-6797 and squashes the following commits:
ca63c86 [Sun Rui] Adjust MimaExcludes after rebase.
7313374 [Sun Rui] Fix unit test errors.
72695fb [Sun Rui] Fix unit test failures.
193882f [Sun Rui] Fix Mima test error.
fe25a33 [Sun Rui] Fix Mima test error.
35ecfa3 [Sun Rui] Fix comments.
c38a005 [Sun Rui] Unzipped SparkR binary package is still required for standalone and Mesos modes.
b05340c [Sun Rui] Fix scala style.
2ca5048 [Sun Rui] Fix comments.
1acefd1 [Sun Rui] Fix scala style.
0aa1e97 [Sun Rui] Fix scala style.
41d4f17 [Sun Rui] Add support for locating SparkR package for R workers required by RDD APIs.
49ff948 [Sun Rui] Invoke jar.exe with full path in install-dev.bat.
7b916c5 [Sun Rui] Use 'rem' consistently.
3bed438 [Sun Rui] Add a comment.
681afb0 [Sun Rui] Fix a bug that RRunner does not handle client deployment modes.
cedfbe2 [Sun Rui] [SPARK-6797][SPARKR] Add support for YARN cluster mode.
This PR adds `metadata` to `InputInfo`. `InputDStream` can report its metadata for a batch and it will be shown in the batch page.
For example,
FileInputDStream will display the new files for a batch, and DirectKafkaInputDStream will display its offset ranges.
Author: zsxwing <zsxwing@gmail.com>
Closes#7081 from zsxwing/input-metadata and squashes the following commits:
f7abd9b [zsxwing] Revert the space changes in project/MimaExcludes.scala
d906209 [zsxwing] Merge branch 'master' into input-metadata
74762da [zsxwing] Fix MiMa tests
7903e33 [zsxwing] Merge branch 'master' into input-metadata
450a46c [zsxwing] Address comments
1d94582 [zsxwing] Raname InputInfo to StreamInputInfo and change "metadata" to Map[String, Any]
d496ae9 [zsxwing] Add input metadata in the batch page
This PR fixes the converter for Python DataFrame, especially for DecimalType
Closes#7106
Author: Davies Liu <davies@databricks.com>
Closes#7131 from davies/decimal_python and squashes the following commits:
4d3c234 [Davies Liu] Merge branch 'master' of github.com:apache/spark into decimal_python
20531d6 [Davies Liu] Merge branch 'master' of github.com:apache/spark into decimal_python
7d73168 [Davies Liu] fix conflit
6cdd86a [Davies Liu] Merge branch 'master' of github.com:apache/spark into decimal_python
7104e97 [Davies Liu] improve type infer
9cd5a21 [Davies Liu] run python tests with SPARK_PREPEND_CLASSES
829a05b [Davies Liu] fix UDT in python
c99e8c5 [Davies Liu] fix mima
c46814a [Davies Liu] convert decimal for Python DataFrames
As rxin suggested in #7298 , we should consider to remove `RDDApi`.
Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>
Closes#7302 from sarutak/remove-rddapi and squashes the following commits:
e495d35 [Kousuke Saruta] Fixed mima
cb7ebb9 [Kousuke Saruta] Removed overriding RDDApi
This PR is a follow-up of #6617 and is part of [SPARK-6774] [2], which aims to ensure interoperability and backwards-compatibility for Spark SQL Parquet support. And this one fixes the read path. Now Spark SQL is expected to be able to read legacy Parquet data files generated by most (if not all) common libraries/tools like parquet-thrift, parquet-avro, and parquet-hive. However, we still need to refactor the write path to write standard Parquet LISTs and MAPs ([SPARK-8848] [4]).
### Major changes
1. `CatalystConverter` class hierarchy refactoring
- Replaces `CatalystConverter` trait with a much simpler `ParentContainerUpdater`.
Now instead of extending the original `CatalystConverter` trait, every converter class accepts an updater which is responsible for propagating the converted value to some parent container. For example, appending array elements to a parent array buffer, appending a key-value pairs to a parent mutable map, or setting a converted value to some specific field of a parent row. Root converter doesn't have a parent and thus uses a `NoopUpdater`.
This simplifies the design since converters don't need to care about details of their parent converters anymore.
- Unifies `CatalystRootConverter`, `CatalystGroupConverter` and `CatalystPrimitiveRowConverter` into `CatalystRowConverter`
Specifically, now all row objects are represented by `SpecificMutableRow` during conversion.
- Refactors `CatalystArrayConverter`, and removes `CatalystArrayContainsNullConverter` and `CatalystNativeArrayConverter`
`CatalystNativeArrayConverter` was probably designed with the intention of avoiding boxing costs. However, the way it uses Scala generics actually doesn't achieve this goal.
The new `CatalystArrayConverter` handles both nullable and non-nullable array elements in a consistent way.
- Implements backwards-compatibility rules in `CatalystArrayConverter`
When Parquet records are being converted, schema of Parquet files should have already been verified. So we only need to care about the structure rather than field names in the Parquet schema. Since all map objects represented in legacy systems have the same structure as the standard one (see [backwards-compatibility rules for MAP] [1]), we only need to deal with LIST (namely array) in `CatalystArrayConverter`.
2. Requested columns handling
When specifying requested columns in `RowReadSupport`, we used to use a Parquet `MessageType` converted from a Catalyst `StructType` which contains all requested columns. This is not preferable when taking compatibility and interoperability into consideration. Because the actual Parquet file may have different physical structure from the converted schema.
In this PR, the schema for requested columns is constructed using the following method:
- For a column that exists in the target Parquet file, we extract the column type by name from the full file schema, and construct a single-field `MessageType` for that column.
- For a column that doesn't exist in the target Parquet file, we create a single-field `StructType` and convert it to a `MessageType` using `CatalystSchemaConverter`.
- Unions all single-field `MessageType`s into a full schema containing all requested fields
With this change, we also fix [SPARK-6123] [3] by validating the global schema against each individual Parquet part-files.
### Testing
This PR also adds compatibility tests for parquet-avro, parquet-thrift, and parquet-hive. Please refer to `README.md` under `sql/core/src/test` for more information about these tests. To avoid build time code generation and adding extra complexity to the build system, Java code generated from testing Thrift schema and Avro IDL is also checked in.
[1]: https://github.com/apache/incubator-parquet-format/blob/master/LogicalTypes.md#backward-compatibility-rules-1
[2]: https://issues.apache.org/jira/browse/SPARK-6774
[3]: https://issues.apache.org/jira/browse/SPARK-6123
[4]: https://issues.apache.org/jira/browse/SPARK-8848
Author: Cheng Lian <lian@databricks.com>
Closes#7231 from liancheng/spark-6776 and squashes the following commits:
360fe18 [Cheng Lian] Adds ParquetHiveCompatibilitySuite
c6fbc06 [Cheng Lian] Removes WIP file committed by mistake
b8c1295 [Cheng Lian] Excludes the whole parquet package from MiMa
598c3e8 [Cheng Lian] Adds extra Maven repo for hadoop-lzo, which is a transitive dependency of parquet-thrift
926af87 [Cheng Lian] Simplifies Parquet compatibility test suites
7946ee1 [Cheng Lian] Fixes Scala styling issues
3d7ab36 [Cheng Lian] Fixes .rat-excludes
a8f13bb [Cheng Lian] Using Parquet writer API to do compatibility tests
f2208cd [Cheng Lian] Adds README.md for Thrift/Avro code generation
1d390aa [Cheng Lian] Adds parquet-thrift compatibility test
440f7b3 [Cheng Lian] Adds generated files to .rat-excludes
13b9121 [Cheng Lian] Adds ParquetAvroCompatibilitySuite
06cfe9d [Cheng Lian] Adds comments about TimestampType handling
a099d3e [Cheng Lian] More comments
0cc1b37 [Cheng Lian] Fixes MiMa checks
884d3e6 [Cheng Lian] Fixes styling issue and reverts unnecessary changes
802cbd7 [Cheng Lian] Fixes bugs related to schema merging and empty requested columns
38fe1e7 [Cheng Lian] Adds explicit return type
7fb21f1 [Cheng Lian] Reverts an unnecessary debugging change
1781dff [Cheng Lian] Adds test case for SPARK-8811
6437d4b [Cheng Lian] Assembles requested schema from Parquet file schema
bcac49f [Cheng Lian] Removes the 16-byte restriction of decimals
a74fb2c [Cheng Lian] More comments
0525346 [Cheng Lian] Removes old Parquet record converters
03c3bd9 [Cheng Lian] Refactors Parquet read path to implement backwards-compatibility rules
All compressed sensing applications, and some of the regression use-cases will have better result by turning the feature scaling off. However, if we implement this naively by training the dataset without doing any standardization, the rate of convergency will not be good. This can be implemented by still standardizing the training dataset but we penalize each component differently to get effectively the same objective function but a better numerical problem. As a result, for those columns with high variances, they will be penalized less, and vice versa. Without this, since all the features are standardized, so they will be penalized the same.
In R, there is an option for this.
`standardize`
Logical flag for x variable standardization, prior to fitting the model sequence. The coefficients are always returned on the original scale. Default is standardize=TRUE. If variables are in the same units already, you might not wish to standardize. See details below for y standardization with family="gaussian".
+cc holdenk mengxr jkbradley
Author: DB Tsai <dbt@netflix.com>
Closes#7080 from dbtsai/lors and squashes the following commits:
877e6c7 [DB Tsai] repahse the doc
7cf45f2 [DB Tsai] address feedback
78d75c9 [DB Tsai] small change
c2c9e60 [DB Tsai] style
6e1a8e0 [DB Tsai] first commit
Matrices allow zeros to be stored in values. Sometimes a method is handy to check if the numNonZeros are same as number of Active values.
Author: MechCoder <manojkumarsivaraj334@gmail.com>
Closes#6904 from MechCoder/nnz_matrix and squashes the following commits:
252c6b7 [MechCoder] Add to MiMa excludes
e2390f5 [MechCoder] Use count instead of foreach
2f62b2f [MechCoder] Add to MiMa excludes
d6e96ef [MechCoder] [SPARK-8479] Add numNonzeros and numActives to linalg.Matrices
This PR introduces `CatalystSchemaConverter` for converting Parquet schema to Spark SQL schema and vice versa. Original conversion code in `ParquetTypesConverter` is removed. Benefits of the new version are:
1. When converting Spark SQL schemas, it generates standard Parquet schemas conforming to [the most updated Parquet format spec] [1]. Converting to old style Parquet schemas is also supported via feature flag `spark.sql.parquet.followParquetFormatSpec` (which is set to `false` for now, and should be set to `true` after both read and write paths are fixed).
Note that although this version of Parquet format spec hasn't been officially release yet, Parquet MR 1.7.0 already sticks to it. So it should be safe to follow.
1. It implements backwards-compatibility rules described in the most updated Parquet format spec. Thus can recognize more schema patterns generated by other/legacy systems/tools.
1. Code organization follows convention used in [parquet-mr] [2], which is easier to follow. (Structure of `CatalystSchemaConverter` is similar to `AvroSchemaConverter`).
To fully implement backwards-compatibility rules in both read and write path, we also need to update `CatalystRowConverter` (which is responsible for converting Parquet records to `Row`s), `RowReadSupport`, and `RowWriteSupport`. These would be done in follow-up PRs.
TODO
- [x] More schema conversion test cases for legacy schema patterns.
[1]: ea09522659/LogicalTypes.md
[2]: https://github.com/apache/parquet-mr/
Author: Cheng Lian <lian@databricks.com>
Closes#6617 from liancheng/spark-6777 and squashes the following commits:
2a2062d [Cheng Lian] Don't convert decimals without precision information
b60979b [Cheng Lian] Adds a constructor which accepts a Configuration, and fixes default value of assumeBinaryIsString
743730f [Cheng Lian] Decimal scale shouldn't be larger than precision
a104a9e [Cheng Lian] Fixes Scala style issue
1f71d8d [Cheng Lian] Adds feature flag to allow falling back to old style Parquet schema conversion
ba84f4b [Cheng Lian] Fixes MapType schema conversion bug
13cb8d5 [Cheng Lian] Fixes MiMa failure
81de5b0 [Cheng Lian] Fixes UDT, workaround read path, and add tests
28ef95b [Cheng Lian] More AnalysisExceptions
b10c322 [Cheng Lian] Replaces require() with analysisRequire() which throws AnalysisException
cceaf3f [Cheng Lian] Implements backwards compatibility rules in CatalystSchemaConverter
Author: Holden Karau <holden@pigscanfly.ca>
Closes#6927 from holdenk/SPARK-7888-Be-able-to-disable-intercept-in-Linear-Regression-in-ML-package and squashes the following commits:
0ad384c [Holden Karau] Add MiMa excludes
4016fac [Holden Karau] Switch to wild card import, remove extra blank lines
ae5baa8 [Holden Karau] CR feedback, move the fitIntercept down rather than changing ymean and etc above
f34971c [Holden Karau] Fix some more long lines
319bd3f [Holden Karau] Fix long lines
3bb9ee1 [Holden Karau] Update the regression suite tests
7015b9f [Holden Karau] Our code performs the same with R, except we need more than one data point but that seems reasonable
0b0c8c0 [Holden Karau] fix the issue with the sample R code
e2140ba [Holden Karau] Add a test, it fails!
5e84a0b [Holden Karau] Write out thoughts and use the correct trait
91ffc0a [Holden Karau] more murh
006246c [Holden Karau] murp?
…ed KafkaRDD methods. Possible fix for [SPARK-7122], but probably a worthwhile optimization regardless.
Author: cody koeninger <cody@koeninger.org>
Closes#6632 from koeninger/kafka-rdd-count and squashes the following commits:
321340d [cody koeninger] [SPARK-8127][Streaming][Kafka] additional test of ordering of take()
5a05d0f [cody koeninger] [SPARK-8127][Streaming][Kafka] additional test of isEmpty
f68bd32 [cody koeninger] [Streaming][Kafka][SPARK-8127] code cleanup
9555b73 [cody koeninger] Merge branch 'master' into kafka-rdd-count
253031d [cody koeninger] [Streaming][Kafka][SPARK-8127] mima exclusion for change to private method
8974b9e [cody koeninger] [Streaming][Kafka][SPARK-8127] check offset ranges before constructing KafkaRDD
c3768c5 [cody koeninger] [Streaming][Kafka] Take advantage of offset range info for size-related KafkaRDD methods. Possible fix for [SPARK-7122], but probably a worthwhile optimization regardless.
see if requiring producer acks eliminates the need for waitUntilLeaderOffset calls in tests
Author: cody koeninger <cody@koeninger.org>
Closes#5921 from koeninger/kafka-0.8.2-test-cleanup and squashes the following commits:
1e89dc8 [cody koeninger] Merge branch 'master' into kafka-0.8.2-test-cleanup
4662828 [cody koeninger] [Streaming][Kafka] filter mima issue for removal of method from private test class
af1e083 [cody koeninger] Merge branch 'master' into kafka-0.8.2-test-cleanup
4298ac2 [cody koeninger] [Streaming][Kafka] update comment to trigger jenkins attempt
1274afb [cody koeninger] [Streaming][Kafka] see if requiring producer acks eliminates the need for waitUntilLeaderOffset calls in tests
This patch replaces Distinct with Aggregate in the optimizer, so Distinct will become
more efficient over time as we optimize Aggregate (via Tungsten).
Author: Reynold Xin <rxin@databricks.com>
Closes#6637 from rxin/replace-distinct and squashes the following commits:
b3cc50e [Reynold Xin] Mima excludes.
93d6117 [Reynold Xin] Code review feedback.
87e4741 [Reynold Xin] [SPARK-7440][SQL] Remove physical Distinct operator in favor of Aggregate.
Author: Patrick Wendell <patrick@databricks.com>
Closes#6328 from pwendell/spark-1.5-update and squashes the following commits:
2f42d02 [Patrick Wendell] A few more excludes
4bebcf0 [Patrick Wendell] Update to RC4
61aaf46 [Patrick Wendell] Using new release candidate
55f1610 [Patrick Wendell] Another exclude
04b4f04 [Patrick Wendell] More issues with transient 1.4 changes
36f549b [Patrick Wendell] [SPARK-7801] [BUILD] Updating versions to SPARK 1.5.0
Author: Holden Karau <holden@pigscanfly.ca>
Closes#6464 from holdenk/SPARK-7910-expose-partitioner-information-in-javardd and squashes the following commits:
de1e644 [Holden Karau] Fix the test to get the partitioner
bdb31cc [Holden Karau] Add Mima exclude for the new method
347ef4c [Holden Karau] Add a quick little test for the partitioner JavaAPI
f49dca9 [Holden Karau] Add partitoner information to JavaRDDLike and fix some whitespace
https://issues.apache.org/jira/browse/SPARK-7805
Because `sql/hive`'s tests depend on the test jar of `sql/core`, we do not need to store `SQLTestUtils` and `ParquetTest` in `src/main`. We should only add stuff that will be needed by `sql/console` or Python tests (for Python, we need it in `src/main`, right? davies).
Author: Yin Huai <yhuai@databricks.com>
Closes#6334 from yhuai/SPARK-7805 and squashes the following commits:
af6d0c9 [Yin Huai] mima
b86746a [Yin Huai] Move SQLTestUtils.scala and ParquetTest.scala to src/test.
There excludes are unnecessary for 1.3 because the changes were made in 1.4.x.
Author: Xiangrui Meng <meng@databricks.com>
Closes#6254 from mengxr/SPARK-7681-mima and squashes the following commits:
7f0cea0 [Xiangrui Meng] remove mima excludes for 1.3
JIRA: https://issues.apache.org/jira/browse/SPARK-7681
Author: Liang-Chi Hsieh <viirya@gmail.com>
Closes#6209 from viirya/sparsevector_gemv and squashes the following commits:
ce0bb8b [Liang-Chi Hsieh] Still need to scal y when beta is 0.0 because it clears out y.
b890e63 [Liang-Chi Hsieh] Do not delete multiply for DenseVector.
57a8c1e [Liang-Chi Hsieh] Add MimaExcludes for v1.4.
458d1ae [Liang-Chi Hsieh] List DenseMatrix.multiply and SparseMatrix.multiply to MimaExcludes too.
054f05d [Liang-Chi Hsieh] Fix scala style.
410381a [Liang-Chi Hsieh] Address comments. Make Matrix.multiply more generalized.
4616696 [Liang-Chi Hsieh] Add support for SparseVector with SparseMatrix.
5d6d07a [Liang-Chi Hsieh] Merge remote-tracking branch 'upstream/master' into sparsevector_gemv
c069507 [Liang-Chi Hsieh] Add SparseVector support for gemv with DenseMatrix.
Replace the DriverQuirks with JdbcDialect(s) (and MySQLDialect/PostgresDialect)
and allow developers to change the dialects on the fly (for new JDBCRRDs only).
Some types (like an unsigned 64bit number) can be trivially mapped to java.
The status quo is that the RRD will fail to load.
This patch makes it possible to overwrite the type mapping to read e.g.
64Bit numbers as strings and handle them afterwards in software.
JDBCSuite has an example that maps all types to String, which should always
work (at the cost of extra code afterwards).
As a side effect it should now be possible to develop simple dialects
out-of-tree and even with spark-shell.
Author: Rene Treffer <treffer@measite.de>
Closes#5555 from rtreffer/jdbc-dialects and squashes the following commits:
3cbafd7 [Rene Treffer] [SPARK-6888] ignore classes belonging to changed API in MIMA report
fe7e2e8 [Rene Treffer] [SPARK-6888] Make the jdbc driver handling user-definable
This patch introduces a new shuffle manager that enhances the existing sort-based shuffle with a new cache-friendly sort algorithm that operates directly on binary data. The goals of this patch are to lower memory usage and Java object overheads during shuffle and to speed up sorting. It also lays groundwork for follow-up patches that will enable end-to-end processing of serialized records.
The new shuffle manager, `UnsafeShuffleManager`, can be enabled by setting `spark.shuffle.manager=tungsten-sort` in SparkConf.
The new shuffle manager uses directly-managed memory to implement several performance optimizations for certain types of shuffles. In cases where the new performance optimizations cannot be applied, the new shuffle manager delegates to SortShuffleManager to handle those shuffles.
UnsafeShuffleManager's optimizations will apply when _all_ of the following conditions hold:
- The shuffle dependency specifies no aggregation or output ordering.
- The shuffle serializer supports relocation of serialized values (this is currently supported
by KryoSerializer and Spark SQL's custom serializers).
- The shuffle produces fewer than 16777216 output partitions.
- No individual record is larger than 128 MB when serialized.
In addition, extra spill-merging optimizations are automatically applied when the shuffle compression codec supports concatenation of serialized streams. This is currently supported by Spark's LZF serializer.
At a high-level, UnsafeShuffleManager's design is similar to Spark's existing SortShuffleManager. In sort-based shuffle, incoming records are sorted according to their target partition ids, then written to a single map output file. Reducers fetch contiguous regions of this file in order to read their portion of the map output. In cases where the map output data is too large to fit in memory, sorted subsets of the output can are spilled to disk and those on-disk files are merged to produce the final output file.
UnsafeShuffleManager optimizes this process in several ways:
- Its sort operates on serialized binary data rather than Java objects, which reduces memory consumption and GC overheads. This optimization requires the record serializer to have certain properties to allow serialized records to be re-ordered without requiring deserialization. See SPARK-4550, where this optimization was first proposed and implemented, for more details.
- It uses a specialized cache-efficient sorter (UnsafeShuffleExternalSorter) that sorts arrays of compressed record pointers and partition ids. By using only 8 bytes of space per record in the sorting array, this fits more of the array into cache.
- The spill merging procedure operates on blocks of serialized records that belong to the same partition and does not need to deserialize records during the merge.
- When the spill compression codec supports concatenation of compressed data, the spill merge simply concatenates the serialized and compressed spill partitions to produce the final output partition. This allows efficient data copying methods, like NIO's `transferTo`, to be used and avoids the need to allocate decompression or copying buffers during the merge.
The shuffle read path is unchanged.
This patch is similar to [SPARK-4550](http://issues.apache.org/jira/browse/SPARK-4550) / #4450 but uses a slightly different implementation. The `unsafe`-based implementation featured in this patch lays the groundwork for followup patches that will enable sorting to operate on serialized data pages that will be prepared by Spark SQL's new `unsafe` operators (such as the new aggregation operator introduced in #5725).
### Future work
There are several tasks that build upon this patch, which will be left to future work:
- [SPARK-7271](https://issues.apache.org/jira/browse/SPARK-7271) Redesign / extend the shuffle interfaces to accept binary data as input. The goal here is to let us bypass serialization steps in cases where the sort input is produced by an operator that operates directly on binary data.
- Extension / redesign of the `Serializer` API. We can add new methods which allow serializers to determine the size requirements for serializing objects and for serializing objects directly to a specified memory address (similar to how `UnsafeRowConverter` works in Spark SQL).
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/apache/spark/5868)
<!-- Reviewable:end -->
Author: Josh Rosen <joshrosen@databricks.com>
Closes#5868 from JoshRosen/unsafe-sort and squashes the following commits:
ef0a86e [Josh Rosen] Fix scalastyle errors
7610f2f [Josh Rosen] Add tests for proper cleanup of shuffle data.
d494ffe [Josh Rosen] Fix deserialization of JavaSerializer instances.
52a9981 [Josh Rosen] Fix some bugs in the address packing code.
51812a7 [Josh Rosen] Change shuffle manager sort name to tungsten-sort
4023fa4 [Josh Rosen] Add @Private annotation to some Java classes.
de40b9d [Josh Rosen] More comments to try to explain metrics code
df07699 [Josh Rosen] Attempt to clarify confusing metrics update code
5e189c6 [Josh Rosen] Track time spend closing / flushing files; split TimeTrackingOutputStream into separate file.
d5779c6 [Josh Rosen] Merge remote-tracking branch 'origin/master' into unsafe-sort
c2ce78e [Josh Rosen] Fix a missed usage of MAX_PARTITION_ID
e3b8855 [Josh Rosen] Cleanup in UnsafeShuffleWriter
4a2c785 [Josh Rosen] rename 'sort buffer' to 'pointer array'
6276168 [Josh Rosen] Remove ability to disable spilling in UnsafeShuffleExternalSorter.
57312c9 [Josh Rosen] Clarify fileBufferSize units
2d4e4f4 [Josh Rosen] Address some minor comments in UnsafeShuffleExternalSorter.
fdcac08 [Josh Rosen] Guard against overflow when expanding sort buffer.
85da63f [Josh Rosen] Cleanup in UnsafeShuffleSorterIterator.
0ad34da [Josh Rosen] Fix off-by-one in nextInt() call
56781a1 [Josh Rosen] Rename UnsafeShuffleSorter to UnsafeShuffleInMemorySorter
e995d1a [Josh Rosen] Introduce MAX_SHUFFLE_OUTPUT_PARTITIONS.
e58a6b4 [Josh Rosen] Add more tests for PackedRecordPointer encoding.
4f0b770 [Josh Rosen] Attempt to implement proper shuffle write metrics.
d4e6d89 [Josh Rosen] Update to bit shifting constants
69d5899 [Josh Rosen] Remove some unnecessary override vals
8531286 [Josh Rosen] Add tests that automatically trigger spills.
7c953f9 [Josh Rosen] Add test that covers UnsafeShuffleSortDataFormat.swap().
e1855e5 [Josh Rosen] Fix a handful of misc. IntelliJ inspections
39434f9 [Josh Rosen] Avoid integer multiplication overflow in getMemoryUsage (thanks FindBugs!)
1e3ad52 [Josh Rosen] Delete unused ByteBufferOutputStream class.
ea4f85f [Josh Rosen] Roll back an unnecessary change in Spillable.
ae538dc [Josh Rosen] Document UnsafeShuffleManager.
ec6d626 [Josh Rosen] Add notes on maximum # of supported shuffle partitions.
0d4d199 [Josh Rosen] Bump up shuffle.memoryFraction to make tests pass.
b3b1924 [Josh Rosen] Properly implement close() and flush() in DummySerializerInstance.
1ef56c7 [Josh Rosen] Revise compression codec support in merger; test cross product of configurations.
b57c17f [Josh Rosen] Disable some overly-verbose logs that rendered DEBUG useless.
f780fb1 [Josh Rosen] Add test demonstrating which compression codecs support concatenation.
4a01c45 [Josh Rosen] Remove unnecessary log message
27b18b0 [Josh Rosen] That for inserting records AT the max record size.
fcd9a3c [Josh Rosen] Add notes + tests for maximum record / page sizes.
9d1ee7c [Josh Rosen] Fix MiMa excludes for ShuffleWriter change
fd4bb9e [Josh Rosen] Use own ByteBufferOutputStream rather than Kryo's
67d25ba [Josh Rosen] Update Exchange operator's copying logic to account for new shuffle manager
8f5061a [Josh Rosen] Strengthen assertion to check partitioning
01afc74 [Josh Rosen] Actually read data in UnsafeShuffleWriterSuite
1929a74 [Josh Rosen] Update to reflect upstream ShuffleBlockManager -> ShuffleBlockResolver rename.
e8718dd [Josh Rosen] Merge remote-tracking branch 'origin/master' into unsafe-sort
9b7ebed [Josh Rosen] More defensive programming RE: cleaning up spill files and memory after errors
7cd013b [Josh Rosen] Begin refactoring to enable proper tests for spilling.
722849b [Josh Rosen] Add workaround for transferTo() bug in merging code; refactor tests.
9883e30 [Josh Rosen] Merge remote-tracking branch 'origin/master' into unsafe-sort
b95e642 [Josh Rosen] Refactor and document logic that decides when to spill.
1ce1300 [Josh Rosen] More minor cleanup
5e8cf75 [Josh Rosen] More minor cleanup
e67f1ea [Josh Rosen] Remove upper type bound in ShuffleWriter interface.
cfe0ec4 [Josh Rosen] Address a number of minor review comments:
8a6fe52 [Josh Rosen] Rename UnsafeShuffleSpillWriter to UnsafeShuffleExternalSorter
11feeb6 [Josh Rosen] Update TODOs related to shuffle write metrics.
b674412 [Josh Rosen] Merge remote-tracking branch 'origin/master' into unsafe-sort
aaea17b [Josh Rosen] Add comments to UnsafeShuffleSpillWriter.
4f70141 [Josh Rosen] Fix merging; now passes UnsafeShuffleSuite tests.
133c8c9 [Josh Rosen] WIP towards testing UnsafeShuffleWriter.
f480fb2 [Josh Rosen] WIP in mega-refactoring towards shuffle-specific sort.
57f1ec0 [Josh Rosen] WIP towards packed record pointers for use in optimized shuffle sort.
69232fd [Josh Rosen] Enable compressible address encoding for off-heap mode.
7ee918e [Josh Rosen] Re-order imports in tests
3aeaff7 [Josh Rosen] More refactoring and cleanup; begin cleaning iterator interfaces
3490512 [Josh Rosen] Misc. cleanup
f156a8f [Josh Rosen] Hacky metrics integration; refactor some interfaces.
2776aca [Josh Rosen] First passing test for ExternalSorter.
5e100b2 [Josh Rosen] Super-messy WIP on external sort
595923a [Josh Rosen] Remove some unused variables.
8958584 [Josh Rosen] Fix bug in calculating free space in current page.
f17fa8f [Josh Rosen] Add missing newline
c2fca17 [Josh Rosen] Small refactoring of SerializerPropertiesSuite to enable test re-use:
b8a09fe [Josh Rosen] Back out accidental log4j.properties change
bfc12d3 [Josh Rosen] Add tests for serializer relocation property.
240864c [Josh Rosen] Remove PrefixComputer and require prefix to be specified as part of insert()
1433b42 [Josh Rosen] Store record length as int instead of long.
026b497 [Josh Rosen] Re-use a buffer in UnsafeShuffleWriter
0748458 [Josh Rosen] Port UnsafeShuffleWriter to Java.
87e721b [Josh Rosen] Renaming and comments
d3cc310 [Josh Rosen] Flag that SparkSqlSerializer2 supports relocation
e2d96ca [Josh Rosen] Expand serializer API and use new function to help control when new UnsafeShuffle path is used.
e267cee [Josh Rosen] Fix compilation of UnsafeSorterSuite
9c6cf58 [Josh Rosen] Refactor to use DiskBlockObjectWriter.
253f13e [Josh Rosen] More cleanup
8e3ec20 [Josh Rosen] Begin code cleanup.
4d2f5e1 [Josh Rosen] WIP
3db12de [Josh Rosen] Minor simplification and sanity checks in UnsafeSorter
767d3ca [Josh Rosen] Fix invalid range in UnsafeSorter.
e900152 [Josh Rosen] Add test for empty iterator in UnsafeSorter
57a4ea0 [Josh Rosen] Make initialSize configurable in UnsafeSorter
abf7bfe [Josh Rosen] Add basic test case.
81d52c5 [Josh Rosen] WIP on UnsafeSorter
JavaTypeInference into catalyst
types.DateUtils into catalyst
CacheManager into execution
DefaultParserDialect into catalyst
Author: Reynold Xin <rxin@databricks.com>
Closes#6108 from rxin/sql-rename and squashes the following commits:
3fc9613 [Reynold Xin] Fixed import ordering.
83d9ff4 [Reynold Xin] Fixed codegen tests.
e271e86 [Reynold Xin] mima
f4e24a6 [Reynold Xin] [SQL] Move some classes into packages that are more appropriate.
This PR migrates Parquet data source to the newly introduced `FSBasedRelation`. `FSBasedParquetRelation` is created to replace `ParquetRelation2`. Major differences are:
1. Partition discovery code has been factored out to `FSBasedRelation`
1. `AppendingParquetOutputFormat` is not used now. Instead, an anonymous subclass of `ParquetOutputFormat` is used to handle appending and writing dynamic partitions
1. When scanning partitioned tables, `FSBasedParquetRelation.buildScan` only builds an `RDD[Row]` for a single selected partition
1. `FSBasedParquetRelation` doesn't rely on Catalyst expressions for filter push down, thus it doesn't extend `CatalystScan` anymore
After migrating `JSONRelation` (which extends `CatalystScan`), we can remove `CatalystScan`.
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/apache/spark/6090)
<!-- Reviewable:end -->
Author: Cheng Lian <lian@databricks.com>
Closes#6090 from liancheng/parquet-migration and squashes the following commits:
6063f87 [Cheng Lian] Casts to OutputCommitter rather than FileOutputCommtter
bfd1cf0 [Cheng Lian] Fixes compilation error introduced while rebasing
f9ea56e [Cheng Lian] Adds ParquetRelation2 related classes to MiMa check whitelist
261d8c1 [Cheng Lian] Minor bug fix and more tests
db65660 [Cheng Lian] Migrates Parquet data source to FSBasedRelation
This PR adds partitioning support for the external data sources API. It aims to simplify development of file system based data sources, and provide first class partitioning support for both read path and write path. Existing data sources like JSON and Parquet can be simplified with this work.
## New features provided
1. Hive compatible partition discovery
This actually generalizes the partition discovery strategy used in Parquet data source in Spark 1.3.0.
1. Generalized partition pruning optimization
Now partition pruning is handled during physical planning phase. Specific data sources don't need to worry about this harness anymore.
(This also implies that we can remove `CatalystScan` after migrating the Parquet data source, since now we don't need to pass Catalyst expressions to data source implementations.)
1. Insertion with dynamic partitions
When inserting data to a `FSBasedRelation`, data can be partitioned dynamically by specified partition columns.
## New structures provided
### Developer API
1. `FSBasedRelation`
Base abstract class for file system based data sources.
1. `OutputWriter`
Base abstract class for output row writers, responsible for writing a single row object.
1. `FSBasedRelationProvider`
A new relation provider for `FSBasedRelation` subclasses. Note that data sources extending `FSBasedRelation` don't need to extend `RelationProvider` and `SchemaRelationProvider`.
### User API
New overloaded versions of
1. `DataFrame.save()`
1. `DataFrame.saveAsTable()`
1. `SQLContext.load()`
are provided to allow users to save/load DataFrames with user defined dynamic partition columns.
### Spark SQL query planning
1. `InsertIntoFSBasedRelation`
Used to implement write path for `FSBasedRelation`s.
1. New rules for `FSBasedRelation` in `DataSourceStrategy`
These are added to hook `FSBasedRelation` into physical query plan in read path, and perform partition pruning.
## TODO
- [ ] Use scratch directories when overwriting a table with data selected from itself.
Currently, this is not supported, because the table been overwritten is always deleted before writing any data to it.
- [ ] When inserting with dynamic partition columns, use external sorter to group the data first.
This ensures that we only need to open a single `OutputWriter` at a time. For data sources like Parquet, `OutputWriter`s can be quite memory consuming. One issue is that, this approach breaks the row distribution in the original DataFrame. However, we did't promise to preserve data distribution when writing a DataFrame.
- [x] More tests. Specifically, test cases for
- [x] Self-join
- [x] Loading partitioned relations with a subset of partition columns stored in data files.
- [x] `SQLContext.load()` with user defined dynamic partition columns.
## Parquet data source migration
Parquet data source migration is covered in PR https://github.com/liancheng/spark/pull/6, which is against this PR branch and for preview only. A formal PR need to be made after this one is merged.
Author: Cheng Lian <lian@databricks.com>
Closes#5526 from liancheng/partitioning-support and squashes the following commits:
5351a1b [Cheng Lian] Fixes compilation error introduced while rebasing
1f9b1a5 [Cheng Lian] Tweaks data schema passed to FSBasedRelations
43ba50e [Cheng Lian] Avoids serializing generated projection code
edf49e7 [Cheng Lian] Removed commented stale code block
348a922 [Cheng Lian] Adds projection in FSBasedRelation.buildScan(requiredColumns, inputPaths)
ad4d4de [Cheng Lian] Enables HDFS style globbing
8d12e69 [Cheng Lian] Fixes compilation error
c71ac6c [Cheng Lian] Addresses comments from @marmbrus
7552168 [Cheng Lian] Fixes typo in MimaExclude.scala
0349e09 [Cheng Lian] Fixes compilation error introduced while rebasing
52b0c9b [Cheng Lian] Adjusts project/MimaExclude.scala
c466de6 [Cheng Lian] Addresses comments
bc3f9b4 [Cheng Lian] Uses projection to separate partition columns and data columns while inserting rows
795920a [Cheng Lian] Fixes compilation error after rebasing
0b8cd70 [Cheng Lian] Adds Scala/Catalyst row conversion when writing non-partitioned tables
fa543f3 [Cheng Lian] Addresses comments
5849dd0 [Cheng Lian] Fixes doc typos. Fixes partition discovery refresh.
51be443 [Cheng Lian] Replaces FSBasedRelation.outputCommitterClass with FSBasedRelation.prepareForWrite
c4ed4fe [Cheng Lian] Bug fixes and a new test suite
a29e663 [Cheng Lian] Bug fix: should only pass actuall data files to FSBaseRelation.buildScan
5f423d3 [Cheng Lian] Bug fixes. Lets data source to customize OutputCommitter rather than OutputFormat
54c3d7b [Cheng Lian] Enforces that FileOutputFormat must be used
be0c268 [Cheng Lian] Uses TaskAttempContext rather than Configuration in OutputWriter.init
0bc6ad1 [Cheng Lian] Resorts to new Hadoop API, and now FSBasedRelation can customize output format class
f320766 [Cheng Lian] Adds prepareForWrite() hook, refactored writer containers
422ff4a [Cheng Lian] Fixes style issue
ce52353 [Cheng Lian] Adds new SQLContext.load() overload with user defined dynamic partition columns
8d2ff71 [Cheng Lian] Merges partition columns when reading partitioned relations
ca1805b [Cheng Lian] Removes duplicated partition discovery code in new Parquet
f18dec2 [Cheng Lian] More strict schema checking
b746ab5 [Cheng Lian] More tests
9b487bf [Cheng Lian] Fixes compilation errors introduced while rebasing
ea6c8dd [Cheng Lian] Removes remote debugging stuff
327bb1d [Cheng Lian] Implements partitioning support for data sources API
3c5073a [Cheng Lian] Fixes SaveModes used in test cases
fb5a607 [Cheng Lian] Fixes compilation error
9d17607 [Cheng Lian] Adds the contract that OutputWriter should have zero-arg constructor
5de194a [Cheng Lian] Forgot Apache licence header
95d0b4d [Cheng Lian] Renames PartitionedSchemaRelationProvider to FSBasedRelationProvider
770b5ba [Cheng Lian] Adds tests for FSBasedRelation
3ba9bbf [Cheng Lian] Adds DataFrame.saveAsTable() overrides which support partitioning
1b8231f [Cheng Lian] Renames FSBasedPrunedFilteredScan to FSBasedRelation
aa8ba9a [Cheng Lian] Javadoc fix
012ed2d [Cheng Lian] Adds PartitioningOptions
7dd8dd5 [Cheng Lian] Adds new interfaces and stub methods for data sources API partitioning support
This PR switches Spark SQL's Hive support to use the isolated hive client interface introduced by #5851, instead of directly interacting with the client. By using this isolated client we can now allow users to dynamically configure the version of Hive that they are connecting to by setting `spark.sql.hive.metastore.version` without the need recompile. This also greatly reduces the surface area for our interaction with the hive libraries, hopefully making it easier to support other versions in the future.
Jars for the desired hive version can be configured using `spark.sql.hive.metastore.jars`, which accepts the following options:
- a colon-separated list of jar files or directories for hive and hadoop.
- `builtin` - attempt to discover the jars that were used to load Spark SQL and use those. This
option is only valid when using the execution version of Hive.
- `maven` - download the correct version of hive on demand from maven.
By default, `builtin` is used for Hive 13.
This PR also removes the test step for building against Hive 12, as this will no longer be required to talk to Hive 12 metastores. However, the full removal of the Shim is deferred until a later PR.
Remaining TODOs:
- Remove the Hive Shims and inline code for Hive 13.
- Several HiveCompatibility tests are not yet passing.
- `nullformatCTAS` - As detailed below, we now are handling CTAS parsing ourselves instead of hacking into the Hive semantic analyzer. However, we currently only handle the common cases and not things like CTAS where the null format is specified.
- `combine1` now leaks state about compression somehow, breaking all subsequent tests. As such we currently add it to the blacklist
- `part_inherit_tbl_props` and `part_inherit_tbl_props_with_star` do not work anymore. We are correctly propagating the information
- "load_dyn_part14.*" - These tests pass when run on their own, but fail when run with all other tests. It seems our `RESET` mechanism may not be as robust as it used to be?
Other required changes:
- `CreateTableAsSelect` no longer carries parts of the HiveQL AST with it through the query execution pipeline. Instead, we parse CTAS during the HiveQL conversion and construct a `HiveTable`. The full parsing here is not yet complete as detailed above in the remaining TODOs. Since the operator is Hive specific, it is moved to the hive package.
- `Command` is simplified to be a trait that simply acts as a marker for a LogicalPlan that should be eagerly evaluated.
Author: Michael Armbrust <michael@databricks.com>
Closes#5876 from marmbrus/useIsolatedClient and squashes the following commits:
258d000 [Michael Armbrust] really really correct path handling
e56fd4a [Michael Armbrust] getAbsolutePath
5a259f5 [Michael Armbrust] fix typos
81bb366 [Michael Armbrust] comments from vanzin
5f3945e [Michael Armbrust] Merge remote-tracking branch 'origin/master' into useIsolatedClient
4b5cd41 [Michael Armbrust] yin's comments
f5de7de [Michael Armbrust] cleanup
11e9c72 [Michael Armbrust] better coverage in versions suite
7e8f010 [Michael Armbrust] better error messages and jar handling
e7b3941 [Michael Armbrust] more permisive checking for function registration
da91ba7 [Michael Armbrust] Merge remote-tracking branch 'origin/master' into useIsolatedClient
5fe5894 [Michael Armbrust] fix serialization suite
81711c4 [Michael Armbrust] Initial support for running without maven
1d8ae44 [Michael Armbrust] fix final tests?
1c50813 [Michael Armbrust] more comments
a3bee70 [Michael Armbrust] Merge remote-tracking branch 'origin/master' into useIsolatedClient
a6f5df1 [Michael Armbrust] style
ab07f7e [Michael Armbrust] WIP
4d8bf02 [Michael Armbrust] Remove hive 12 compilation
8843a25 [Michael Armbrust] [SPARK-6908] [SQL] Use isolated Hive client
Now that 1.3 has been released, we should enable MiMa checks for the `sql` subproject.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#5727 from JoshRosen/enable-more-mima-checks and squashes the following commits:
3ad302b [Josh Rosen] Merge remote-tracking branch 'origin/master' into enable-more-mima-checks
0c48e4d [Josh Rosen] Merge remote-tracking branch 'origin/master' into enable-more-mima-checks
e276cee [Josh Rosen] Fix SQL MiMa checks via excludes and private[sql]
44d0d01 [Josh Rosen] Add back 'launcher' exclude
1aae027 [Josh Rosen] Enable MiMa checks for launcher and sql projects.
Add `compressed` to `Vector` with some other methods: `numActives`, `numNonzeros`, `toSparse`, and `toDense`. jkbradley
Author: Xiangrui Meng <meng@databricks.com>
Closes#5756 from mengxr/SPARK-6756 and squashes the following commits:
8d4ecbd [Xiangrui Meng] address comment and add mima excludes
da54179 [Xiangrui Meng] add toSparse, toDense, numActives, numNonzeros, and compressed to Vector
jira: https://issues.apache.org/jira/browse/SPARK-7090
LDA was implemented with extensibility in mind. And with the development of OnlineLDA and Gibbs Sampling, we are collecting more detailed requirements from different algorithms.
As Joseph Bradley jkbradley proposed in https://github.com/apache/spark/pull/4807 and with some further discussion, we'd like to adjust the code structure a little to present the common interface and extension point clearly.
Basically class LDA would be a common entrance for LDA computing. And each LDA object will refer to a LDAOptimizer for the concrete algorithm implementation. Users can customize LDAOptimizer with specific parameters and assign it to LDA.
Concrete changes:
1. Add a trait `LDAOptimizer`, which defines the common iterface for concrete implementations. Each subClass is a wrapper for a specific LDA algorithm.
2. Move EMOptimizer to file LDAOptimizer and inherits from LDAOptimizer, rename to EMLDAOptimizer. (in case a more generic EMOptimizer comes in the future)
-adjust the constructor of EMOptimizer, since all the parameters should be passed in through initialState method. This can avoid unwanted confusion or overwrite.
-move the code from LDA.initalState to initalState of EMLDAOptimizer
3. Add property ldaOptimizer to LDA and its getter/setter, and EMLDAOptimizer is the default Optimizer.
4. Change the return type of LDA.run from DistributedLDAModel to LDAModel.
Further work:
add OnlineLDAOptimizer and other possible Optimizers once ready.
Author: Yuhao Yang <hhbyyh@gmail.com>
Closes#5661 from hhbyyh/ldaRefactor and squashes the following commits:
0e2e006 [Yuhao Yang] respond to review comments
08a45da [Yuhao Yang] Merge remote-tracking branch 'upstream/master' into ldaRefactor
e756ce4 [Yuhao Yang] solve mima exception
d74fd8f [Yuhao Yang] Merge remote-tracking branch 'upstream/master' into ldaRefactor
0bb8400 [Yuhao Yang] refactor LDA with Optimizer
ec2f857 [Yuhao Yang] protoptype for discussion
I've added a static getOrCreate method to the static SparkContext object that allows one to either retrieve a previously created SparkContext or to instantiate a new one with the provided config. The method accepts an optional SparkConf to make usage intuitive.
Still working on a test for this, basically want to create a new context from scratch, then ensure that subsequent calls don't overwrite that.
Author: Ilya Ganelin <ilya.ganelin@capitalone.com>
Closes#5501 from ilganeli/SPARK-6703 and squashes the following commits:
db9a963 [Ilya Ganelin] Closing second spark context
1dc0444 [Ilya Ganelin] Added ref equality check
8c884fa [Ilya Ganelin] Made getOrCreate synchronized
cb0c6b7 [Ilya Ganelin] Doc updates and code cleanup
270cfe3 [Ilya Ganelin] [SPARK-6703] Documentation fixes
15e8dea [Ilya Ganelin] Updated comments and added MiMa Exclude
0e1567c [Ilya Ganelin] Got rid of unecessary option for AtomicReference
dfec4da [Ilya Ganelin] Changed activeContext to AtomicReference
733ec9f [Ilya Ganelin] Fixed some bugs in test code
8be2f83 [Ilya Ganelin] Replaced match with if
e92caf7 [Ilya Ganelin] [SPARK-6703] Added test to ensure that getOrCreate both allows creation, retrieval, and a second context if desired
a99032f [Ilya Ganelin] Spacing fix
d7a06b8 [Ilya Ganelin] Updated SparkConf class to add getOrCreate method. Started test suite implementation
jira: https://issues.apache.org/jira/browse/SPARK-6693
It's kind of annoying when debugging and found you cannot print out the matrix as you want.
original toString of Matrix only print like following,
0.17810102596909183 0.5616906241468385 ... (10 total)
0.9692861997823815 0.015558159784155756 ...
0.8513015122819192 0.031523763918528847 ...
0.5396875653953941 0.3267864552779176 ...
The def toString(maxLines : Int, maxWidth : Int) is useful when debuging, logging and saving matrix to files.
Author: Yuhao Yang <hhbyyh@gmail.com>
Closes#5344 from hhbyyh/addToString and squashes the following commits:
19a6836 [Yuhao Yang] remove extra line
6314b21 [Yuhao Yang] add exclude
736c324 [Yuhao Yang] add ut and exclude
420da39 [Yuhao Yang] Merge remote-tracking branch 'upstream/master' into addToString
c22f352 [Yuhao Yang] style change
64a9e0f [Yuhao Yang] add specific to string to matrix
I've added a timeout and retry loop around the SparkContext shutdown code that should fix this deadlock. If a SparkContext shutdown is in progress when another thread comes knocking, it will wait for 10 seconds for the lock, then fall through where the outer loop will re-submit the request.
Author: Ilya Ganelin <ilya.ganelin@capitalone.com>
Closes#5277 from ilganeli/SPARK-6492 and squashes the following commits:
8617a7e [Ilya Ganelin] Resolved merge conflict
2fbab66 [Ilya Ganelin] Added MIMA Exclude
a0e2c70 [Ilya Ganelin] Deleted stale imports
fa28ce7 [Ilya Ganelin] reverted to just having a single stopped
76fc825 [Ilya Ganelin] Updated to use atomic booleans instead of the synchronized vars
6e8a7f7 [Ilya Ganelin] Removing unecessary null check for now since i'm not fixing stop ordering yet
cdf7073 [Ilya Ganelin] [SPARK-6492] Moved stopped=true back to the start of the shutdown sequence so this can be addressed in a seperate PR
7fb795b [Ilya Ganelin] Spacing
b7a0c5c [Ilya Ganelin] Import ordering
df8224f [Ilya Ganelin] Added comment for added lock
343cb94 [Ilya Ganelin] [SPARK-6492] Added timeout/retry logic to fix a deadlock in SparkContext shutdown
Hi all - this patch changes the Stage class to an abstract class and introduces two new classes that extend it: ShuffleMapStage and ResultStage - with the goal of increasing readability of the DAGScheduler class. Their usage is updated within DAGScheduler.
Author: Ilya Ganelin <ilya.ganelin@capitalone.com>
Author: Ilya Ganelin <ilganeli@gmail.com>
Closes#4708 from ilganeli/SPARK-4655 and squashes the following commits:
c248924 [Ilya Ganelin] Merge branch 'SPARK-4655' of github.com:ilganeli/spark into SPARK-4655
d930385 [Ilya Ganelin] Fixed merge conflict from
a9a765f [Ilya Ganelin] Update DAGScheduler.scala
c03563c [Ilya Ganelin] Minor fixeS
c39e971 [Ilya Ganelin] Added return typing for public methods
845bc87 [Ilya Ganelin] Merge branch 'SPARK-4655' of github.com:ilganeli/spark into SPARK-4655
e8031d8 [Ilya Ganelin] Minor string fixes
4ec53ac [Ilya Ganelin] Merge remote-tracking branch 'upstream/master' into SPARK-4655
c004f62 [Ilya Ganelin] Update DAGScheduler.scala
a2cb03f [Ilya Ganelin] [SPARK-4655] Replaced usages of Nil and eliminated some code reuse
3d5cf20 [Ilya Ganelin] [SPARK-4655] Moved mima exclude to 1.4
6912c55 [Ilya Ganelin] Resolved merge conflict
4bff208 [Ilya Ganelin] Minor stylistic fixes
c6fffbb [Ilya Ganelin] newline
41402ad [Ilya Ganelin] Style fixes
02c6981 [Ilya Ganelin] Merge branch 'SPARK-4655' of github.com:ilganeli/spark into SPARK-4655
c755a09 [Ilya Ganelin] Some more stylistic updates and minor refactoring
b6257a0 [Ilya Ganelin] Update MimaExcludes.scala
0f0c624 [Ilya Ganelin] Fixed merge conflict
2eba262 [Ilya Ganelin] Merge remote-tracking branch 'upstream/master' into SPARK-4655
6b43d7b [Ilya Ganelin] Got rid of some spaces
6f1a5db [Ilya Ganelin] Revert "More minor formatting and refactoring"
1b3471b [Ilya Ganelin] Merge remote-tracking branch 'upstream/master' into SPARK-4655
c9288e2 [Ilya Ganelin] More minor formatting and refactoring
d548caf [Ilya Ganelin] Formatting fix
c3ae5c2 [Ilya Ganelin] Explicit typing
0dacaf3 [Ilya Ganelin] Got rid of stale import
6da3a71 [Ilya Ganelin] Trailing whitespace
b85c5fe [Ilya Ganelin] Added minor fixes
a57dfcd [Ilya Ganelin] Added MiMA exclusion to get around binary compatibility check
83ed849 [Ilya Ganelin] moved braces for consistency
96dd161 [Ilya Ganelin] Fixed minor style error
cfd6f10 [Ilya Ganelin] Updated DAGScheduler to use new ResultStage and ShuffleMapStage classes
83494e9 [Ilya Ganelin] Added new Stage classes
This PR added a standard internal RPC interface for Spark and an Akka implementation. See [the design document](https://issues.apache.org/jira/secure/attachment/12698710/Pluggable%20RPC%20-%20draft%202.pdf) for more details.
I will split the whole work into multiple PRs to make it easier for code review. This is the first PR and avoid to touch too many files.
Author: zsxwing <zsxwing@gmail.com>
Closes#4588 from zsxwing/rpc-part1 and squashes the following commits:
fe3df4c [zsxwing] Move registerEndpoint and use actorSystem.dispatcher in asyncSetupEndpointRefByURI
f6f3287 [zsxwing] Remove RpcEndpointRef.toURI
8bd1097 [zsxwing] Fix docs and the code style
f459380 [zsxwing] Add RpcAddress.fromURI and rename urls to uris
b221398 [zsxwing] Move send methods above ask methods
15cfd7b [zsxwing] Merge branch 'master' into rpc-part1
9ffa997 [zsxwing] Fix MiMa tests
78a1733 [zsxwing] Merge remote-tracking branch 'origin/master' into rpc-part1
385b9c3 [zsxwing] Fix the code style and add docs
2cc3f78 [zsxwing] Add an asynchronous version of setupEndpointRefByUrl
e8dfec3 [zsxwing] Remove 'sendWithReply(message: Any, sender: RpcEndpointRef): Unit'
08564ae [zsxwing] Add RpcEnvFactory to create RpcEnv
e5df4ca [zsxwing] Handle AkkaFailure(e) in Actor
ec7c5b0 [zsxwing] Fix docs
7fc95e1 [zsxwing] Implement askWithReply in RpcEndpointRef
9288406 [zsxwing] Document thread-safety for setupThreadSafeEndpoint
3007c09 [zsxwing] Move setupDriverEndpointRef to RpcUtils and rename to makeDriverRef
c425022 [zsxwing] Fix the code style
5f87700 [zsxwing] Move the logical of processing message to a private function
3e56123 [zsxwing] Use lazy to eliminate CountDownLatch
07f128f [zsxwing] Remove ActionScheduler.scala
4d34191 [zsxwing] Remove scheduler from RpcEnv
7cdd95e [zsxwing] Add docs for RpcEnv
51e6667 [zsxwing] Add 'sender' to RpcCallContext and rename the parameter of receiveAndReply to 'context'
ffc1280 [zsxwing] Rename 'fail' to 'sendFailure' and other minor code style changes
28e6d0f [zsxwing] Add onXXX for network events and remove the companion objects of network events
3751c97 [zsxwing] Rename RpcResponse to RpcCallContext
fe7d1ff [zsxwing] Add explicit reply in rpc
7b9e0c9 [zsxwing] Fix the indentation
04a106e [zsxwing] Remove NopCancellable and add a const NOP in object SettableCancellable
2a579f4 [zsxwing] Remove RpcEnv.systemName
155b987 [zsxwing] Change newURI to uriOf and add some comments
45b2317 [zsxwing] A standard RPC interface and An Akka implementation
Adds a `Graph#minus` method which will return only unique `VertexId`'s from the calling `VertexRDD`.
To demonstrate a basic example with pseudocode:
```
Set((0L,0),(1L,1)).minus(Set((1L,1),(2L,2)))
> Set((0L,0))
```
Author: Brennon York <brennon.york@capitalone.com>
Closes#5175 from brennonyork/SPARK-6510 and squashes the following commits:
248d5c8 [Brennon York] added minus(VertexRDD[VD]) method to avoid createUsingIndex and updated the mask operations to simplify with andNot call
3fb7cce [Brennon York] updated graphx doc to reflect the addition of minus method
6575d92 [Brennon York] updated mima exclude
aaa030b [Brennon York] completed graph#minus functionality
7227c0f [Brennon York] beginning work on minus functionality
Changed method invocation of 'diff' to match that of 'innerJoin' and 'leftJoin' from VertexRDD[VD] to RDD[(VertexId, VD)]. This change maintains backwards compatibility and better unifies the VertexRDD methods to match each other.
Author: Brennon York <brennon.york@capitalone.com>
Closes#4733 from brennonyork/SPARK-5922 and squashes the following commits:
e800f08 [Brennon York] fixed merge conflicts
b9274af [Brennon York] fixed merge conflicts
f86375c [Brennon York] fixed minor include line
398ddb4 [Brennon York] fixed merge conflicts
aac1810 [Brennon York] updated to aggregateUsingIndex and added test to ensure that method works properly
2af0b88 [Brennon York] removed deprecation line
753c963 [Brennon York] fixed merge conflicts and set preference to use the diff(other: VertexRDD[VD]) method
2c678c6 [Brennon York] added mima exclude to exclude new public diff method from VertexRDD
93186f3 [Brennon York] added back the original diff method to sustain binary compatibility
f18356e [Brennon York] changed method invocation of 'diff' to match that of 'innerJoin' and 'leftJoin' from VertexRDD[VD] to RDD[(VertexId, VD)]
The issue is discussed in https://issues.apache.org/jira/browse/SPARK-5669. Replacing all JBLAS usage by netlib-java gives us a simpler dependency tree and less license issues to worry about. I didn't touch the test scope in this PR. The user guide is not modified to avoid merge conflicts with branch-1.3. srowen ankurdave pwendell
Author: Xiangrui Meng <meng@databricks.com>
Closes#4699 from mengxr/SPARK-5814 and squashes the following commits:
48635c6 [Xiangrui Meng] move netlib-java version to parent pom
ca21c74 [Xiangrui Meng] remove jblas from ml-guide
5f7767a [Xiangrui Meng] Merge remote-tracking branch 'apache/master' into SPARK-5814
c5c4183 [Xiangrui Meng] merge master
0f20cad [Xiangrui Meng] add mima excludes
e53e9f4 [Xiangrui Meng] remove jblas from mllib runtime
ceaa14d [Xiangrui Meng] replace jblas by netlib-java in graphx
fa7c2ca [Xiangrui Meng] move jblas to test scope
