The purpose of this PR is to keep the unsafe format detail only inside the unsafe class itself, so when we use them(like use unsafe array in unsafe map, use unsafe array and map in columnar cache), we don't need to understand the format before use them.
change list:
* unsafe array's 4-bytes numElements header is now required(was optional), and become a part of unsafe array format.
* w.r.t the previous changing, the `sizeInBytes` of unsafe array now counts the 4-bytes header.
* unsafe map's format was `[numElements] [key array numBytes] [key array content(without numElements header)] [value array content(without numElements header)]` before, which is a little hacky as it makes unsafe array's header optional. I think saving 4 bytes is not a big deal, so the format is now: `[key array numBytes] [unsafe key array] [unsafe value array]`.
* w.r.t the previous changing, the `sizeInBytes` of unsafe map now counts both map's header and array's header.
Author: Wenchen Fan <wenchen@databricks.com>
Closes#9131 from cloud-fan/unsafe.
Some json parsers are not closed. parser in JacksonParser#parseJson, for example.
Author: navis.ryu <navis@apache.org>
Closes#9130 from navis/SPARK-11124.
Actually all of the `UnaryMathExpression` doens't support the Decimal, will create follow ups for supporing it. This is the first PR which will be good to review the approach I am taking.
Author: Cheng Hao <hao.cheng@intel.com>
Closes#9086 from chenghao-intel/ceiling.
This patch extends TungstenAggregate to support ImperativeAggregate functions. The existing TungstenAggregate operator only supported DeclarativeAggregate functions, which are defined in terms of Catalyst expressions and can be evaluated via generated projections. ImperativeAggregate functions, on the other hand, are evaluated by calling their `initialize`, `update`, `merge`, and `eval` methods.
The basic strategy here is similar to how SortBasedAggregate evaluates both types of aggregate functions: use a generated projection to evaluate the expression-based declarative aggregates with dummy placeholder expressions inserted in place of the imperative aggregate function output, then invoke the imperative aggregate functions and target them against the aggregation buffer. The bulk of the diff here consists of code that was copied and adapted from SortBasedAggregate, with some key changes to handle TungstenAggregate's sort fallback path.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#9038 from JoshRosen/support-interpreted-in-tungsten-agg-final.
Right now, we have QualifiedTableName, TableIdentifier, and Seq[String] to represent table identifiers. We should only have one form and TableIdentifier is the best one because it provides methods to get table name, database name, return unquoted string, and return quoted string.
Author: Wenchen Fan <wenchen@databricks.com>
Author: Wenchen Fan <cloud0fan@163.com>
Closes#8453 from cloud-fan/table-name.
We should not stop resolving having when the having condtion is resolved, or something like `count(1)` will crash.
Author: Wenchen Fan <cloud0fan@163.com>
Closes#9105 from cloud-fan/having.
This is a first draft of the ability to construct expressions that will take a catalyst internal row and construct a Product (case class or tuple) that has fields with the correct names. Support include:
- Nested classes
- Maps
- Efficiently handling of arrays of primitive types
Not yet supported:
- Case classes that require custom collection types (i.e. List instead of Seq).
Author: Michael Armbrust <michael@databricks.com>
Closes#9100 from marmbrus/productContructor.
In the current implementation of named expressions' `ExprIds`, we rely on a per-JVM AtomicLong to ensure that expression ids are unique within a JVM. However, these expression ids will not be _globally_ unique. This opens the potential for id collisions if new expression ids happen to be created inside of tasks rather than on the driver.
There are currently a few cases where tasks allocate expression ids, which happen to be safe because those expressions are never compared to expressions created on the driver. In order to guard against the introduction of invalid comparisons between driver-created and executor-created expression ids, this patch extends `ExprId` to incorporate a UUID to identify the JVM that created the id, which prevents collisions.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#9093 from JoshRosen/SPARK-11080.
This PR improve the unrolling and read of complex types in columnar cache:
1) Using UnsafeProjection to do serialization of complex types, so they will not be serialized three times (two for actualSize)
2) Copy the bytes from UnsafeRow/UnsafeArrayData to ByteBuffer directly, avoiding the immediate byte[]
3) Using the underlying array in ByteBuffer to create UTF8String/UnsafeRow/UnsafeArrayData without copy.
Combine these optimizations, we can reduce the unrolling time from 25s to 21s (20% less), reduce the scanning time from 3.5s to 2.5s (28% less).
```
df = sqlContext.read.parquet(path)
t = time.time()
df.cache()
df.count()
print 'unrolling', time.time() - t
for i in range(10):
t = time.time()
print df.select("*")._jdf.queryExecution().toRdd().count()
print time.time() - t
```
The schema is
```
root
|-- a: struct (nullable = true)
| |-- b: long (nullable = true)
| |-- c: string (nullable = true)
|-- d: array (nullable = true)
| |-- element: long (containsNull = true)
|-- e: map (nullable = true)
| |-- key: long
| |-- value: string (valueContainsNull = true)
```
Now the columnar cache depends on that UnsafeProjection support all the data types (including UDT), this PR also fix that.
Author: Davies Liu <davies@databricks.com>
Closes#9016 from davies/complex2.
JIRA: https://issues.apache.org/jira/browse/SPARK-10960
When accessing a column in inner select from a select with window function, `AnalysisException` will be thrown. For example, an query like this:
select area, rank() over (partition by area order by tmp.month) + tmp.tmp1 as c1 from (select month, area, product, 1 as tmp1 from windowData) tmp
Currently, the rule `ExtractWindowExpressions` in `Analyzer` only extracts regular expressions from `WindowFunction`, `WindowSpecDefinition` and `AggregateExpression`. We need to also extract other attributes as the one in `Alias` as shown in the above query.
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes#9011 from viirya/fix-window-inner-column.
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.
UnsafeRow contains 3 pieces of information when pointing to some data in memory (an object, a base offset, and length). When the row is serialized with Java/Kryo serialization, the object layout in memory can change if two machines have different pointer width (Oops in JVM).
To reproduce, launch Spark using
MASTER=local-cluster[2,1,1024] bin/spark-shell --conf "spark.executor.extraJavaOptions=-XX:-UseCompressedOops"
And then run the following
scala> sql("select 1 xx").collect()
Author: Reynold Xin <rxin@databricks.com>
Closes#9030 from rxin/SPARK-10914.
This PR refactors Parquet write path to follow parquet-format spec. It's a successor of PR #7679, but with less non-essential changes.
Major changes include:
1. Replaces `RowWriteSupport` and `MutableRowWriteSupport` with `CatalystWriteSupport`
- Writes Parquet data using standard layout defined in parquet-format
Specifically, we are now writing ...
- ... arrays and maps in standard 3-level structure with proper annotations and field names
- ... decimals as `INT32` and `INT64` whenever possible, and taking `FIXED_LEN_BYTE_ARRAY` as the final fallback
- Supports legacy mode which is compatible with Spark 1.4 and prior versions
The legacy mode is by default off, and can be turned on by flipping SQL option `spark.sql.parquet.writeLegacyFormat` to `true`.
- Eliminates per value data type dispatching costs via prebuilt composed writer functions
1. Cleans up the last pieces of old Parquet support code
As pointed out by rxin previously, we probably want to rename all those `Catalyst*` Parquet classes to `Parquet*` for clarity. But I'd like to do this in a follow-up PR to minimize code review noises in this one.
Author: Cheng Lian <lian@databricks.com>
Closes#8988 from liancheng/spark-8848/standard-parquet-write-path.
This PR is a first cut at code generating an encoder that takes a Scala `Product` type and converts it directly into the tungsten binary format. This is done through the addition of a new set of expression that can be used to invoke methods on raw JVM objects, extracting fields and converting the result into the required format. These can then be used directly in an `UnsafeProjection` allowing us to leverage the existing encoding logic.
According to some simple benchmarks, this can significantly speed up conversion (~4x). However, replacing CatalystConverters is deferred to a later PR to keep this PR at a reasonable size.
```scala
case class SomeInts(a: Int, b: Int, c: Int, d: Int, e: Int)
val data = SomeInts(1, 2, 3, 4, 5)
val encoder = ProductEncoder[SomeInts]
val converter = CatalystTypeConverters.createToCatalystConverter(ScalaReflection.schemaFor[SomeInts].dataType)
(1 to 5).foreach {iter =>
benchmark(s"converter $iter") {
var i = 100000000
while (i > 0) {
val res = converter(data).asInstanceOf[InternalRow]
assert(res.getInt(0) == 1)
assert(res.getInt(1) == 2)
i -= 1
}
}
benchmark(s"encoder $iter") {
var i = 100000000
while (i > 0) {
val res = encoder.toRow(data)
assert(res.getInt(0) == 1)
assert(res.getInt(1) == 2)
i -= 1
}
}
}
```
Results:
```
[info] converter 1: 7170ms
[info] encoder 1: 1888ms
[info] converter 2: 6763ms
[info] encoder 2: 1824ms
[info] converter 3: 6912ms
[info] encoder 3: 1802ms
[info] converter 4: 7131ms
[info] encoder 4: 1798ms
[info] converter 5: 7350ms
[info] encoder 5: 1912ms
```
Author: Michael Armbrust <michael@databricks.com>
Closes#9019 from marmbrus/productEncoder.
This PR refactors `HashJoinNode` to take a existing `HashedRelation`. So, we can reuse this node for both `ShuffledHashJoin` and `BroadcastHashJoin`.
https://issues.apache.org/jira/browse/SPARK-10887
Author: Yin Huai <yhuai@databricks.com>
Closes#8953 from yhuai/SPARK-10887.
In the analysis phase , while processing the rules for IN predicate, we
compare the in-list types to the lhs expression type and generate
cast operation if necessary. In the case of NULL [NOT] IN expr1 , we end up
generating cast between in list types to NULL like cast (1 as NULL) which
is not a valid cast.
The fix is to not generate such a cast if the lhs type is a NullType instead
we translate the expression to Literal(Null).
Author: Dilip Biswal <dbiswal@us.ibm.com>
Closes#8983 from dilipbiswal/spark_8654.
Its pretty hard to debug problems with expressions when you can't see all the arguments.
Before: `invoke()`
After: `invoke(inputObject#1, intField, IntegerType)`
Author: Michael Armbrust <michael@databricks.com>
Closes#9022 from marmbrus/expressionToString.
This PR improve the performance of complex types in columnar cache by using UnsafeProjection instead of KryoSerializer.
A simple benchmark show that this PR could improve the performance of scanning a cached table with complex columns by 15x (comparing to Spark 1.5).
Here is the code used to benchmark:
```
df = sc.range(1<<23).map(lambda i: Row(a=Row(b=i, c=str(i)), d=range(10), e=dict(zip(range(10), [str(i) for i in range(10)])))).toDF()
df.write.parquet("table")
```
```
df = sqlContext.read.parquet("table")
df.cache()
df.count()
t = time.time()
print df.select("*")._jdf.queryExecution().toRdd().count()
print time.time() - t
```
Author: Davies Liu <davies@databricks.com>
Closes#8971 from davies/complex.
This patch allows `Repartition` to support UnsafeRows. This is accomplished by implementing the logical `Repartition` operator in terms of `Exchange` and a new `RoundRobinPartitioning`.
Author: Josh Rosen <joshrosen@databricks.com>
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes#8083 from JoshRosen/SPARK-9702.
The created decimal is wrong if using `Decimal(unscaled, precision, scale)` with unscaled > 1e18 and and precision > 18 and scale > 0.
This bug exists since the beginning.
Author: Davies Liu <davies@databricks.com>
Closes#9014 from davies/fix_decimal.
DeclarativeAggregate matches more closely with ImperativeAggregate we already have.
Author: Reynold Xin <rxin@databricks.com>
Closes#9013 from rxin/SPARK-10982.
This patch refactors several of the Aggregate2 interfaces in order to improve code clarity.
The biggest change is a refactoring of the `AggregateFunction2` class hierarchy. In the old code, we had a class named `AlgebraicAggregate` that inherited from `AggregateFunction2`, added a new set of methods, then banned the use of the inherited methods. I found this to be fairly confusing because.
If you look carefully at the existing code, you'll see that subclasses of `AggregateFunction2` fall into two disjoint categories: imperative aggregation functions which directly extended `AggregateFunction2` and declarative, expression-based aggregate functions which extended `AlgebraicAggregate`. In order to make this more explicit, this patch refactors things so that `AggregateFunction2` is a sealed abstract class with two subclasses, `ImperativeAggregateFunction` and `ExpressionAggregateFunction`. The superclass, `AggregateFunction2`, now only contains methods and fields that are common to both subclasses.
After making this change, I updated the various AggregationIterator classes to comply with this new naming scheme. I also performed several small renamings in the aggregate interfaces themselves in order to improve clarity and rewrote or expanded a number of comments.
Author: Josh Rosen <joshrosen@databricks.com>
Closes#8973 from JoshRosen/tungsten-agg-comments.
This PR is mostly cosmetic and cleans up some warts in codegen (nearly all of which were inherited from the original quasiquote version).
- Add lines numbers to errors (in stacktraces when debug logging is on, and always for compile fails)
- Use a variable for input row instead of hardcoding "i" everywhere
- rename `primitive` -> `value` (since its often actually an object)
Author: Michael Armbrust <michael@databricks.com>
Closes#9006 from marmbrus/codegen-cleanup.
`Murmur3_x86_32.hashUnsafeWords` only accepts word-aligned bytes, but unsafe array is not.
Author: Wenchen Fan <cloud0fan@163.com>
Closes#8987 from cloud-fan/hash.
This PR is a completely rewritten of GenerateUnsafeProjection, to accomplish the goal of copying data only once. The old code of GenerateUnsafeProjection is still there to reduce review difficulty.
Instead of creating unsafe conversion code for struct, array and map, we create code of writing the content to the global row buffer.
Author: Wenchen Fan <cloud0fan@163.com>
Author: Wenchen Fan <cloud0fan@outlook.com>
Closes#8747 from cloud-fan/copy-once.
The utilities such as Substring#substringBinarySQL and BinaryPrefixComparator#computePrefix for binary data are put together in ByteArray for easy-to-read.
Author: Takeshi YAMAMURO <linguin.m.s@gmail.com>
Closes#8122 from maropu/CleanUpForBinaryType.
Floor & Ceiling function should returns Long type, rather than Double.
Verified with MySQL & Hive.
Author: Cheng Hao <hao.cheng@intel.com>
Closes#8933 from chenghao-intel/ceiling.
This is an implementation of Hive's `json_tuple` function using Jackson Streaming.
Author: Nathan Howell <nhowell@godaddy.com>
Closes#7946 from NathanHowell/SPARK-9617.
This PR implements a HyperLogLog based Approximate Count Distinct function using the new UDAF interface.
The implementation is inspired by the ClearSpring HyperLogLog implementation and should produce the same results.
There is still some documentation and testing left to do.
cc yhuai
Author: Herman van Hovell <hvanhovell@questtec.nl>
Closes#8362 from hvanhovell/SPARK-9741.
When reading Parquet string and binary-backed decimal values, Parquet `Binary.getBytes` always returns a copied byte array, which is unnecessary. Since the underlying implementation of `Binary` values there is guaranteed to be `ByteArraySliceBackedBinary`, and Parquet itself never reuses underlying byte arrays, we can use `Binary.toByteBuffer.array()` to steal the underlying byte arrays without copying them.
This brings performance benefits when scanning Parquet string and binary-backed decimal columns. Note that, this trick doesn't cover binary-backed decimals with precision greater than 18.
My micro-benchmark result is that, this brings a ~15% performance boost for scanning TPC-DS `store_sales` table (scale factor 15).
Another minor optimization done in this PR is that, now we directly construct a Java `BigDecimal` in `Decimal.toJavaBigDecimal` without constructing a Scala `BigDecimal` first. This brings another ~5% performance gain.
Author: Cheng Lian <lian@databricks.com>
Closes#8907 from liancheng/spark-10811/eliminate-array-copying.
https://issues.apache.org/jira/browse/SPARK-10741
I choose the second approach: do not change output exprIds when convert MetastoreRelation to LogicalRelation
Author: Wenchen Fan <cloud0fan@163.com>
Closes#8889 from cloud-fan/hot-bug.
Since `scala.util.parsing.combinator.Parsers` is thread-safe since Scala 2.10 (See [SI-4929](https://issues.scala-lang.org/browse/SI-4929)), we can change SqlParser to object to avoid memory leak.
I didn't change other subclasses of `scala.util.parsing.combinator.Parsers` because there is only one instance in one SQLContext, which should not be an issue.
Author: zsxwing <zsxwing@gmail.com>
Closes#8357 from zsxwing/sql-memory-leak.
From JIRA: Schema merging should only handle struct fields. But currently we also reconcile decimal precision and scale information.
Author: Holden Karau <holden@pigscanfly.ca>
Closes#8634 from holdenk/SPARK-10449-dont-merge-different-precision.
Intersect and Except are both set operators and they use the all the columns to compare equality between rows. When pushing their Project parent down, the relations they based on would change, therefore not an equivalent transformation.
JIRA: https://issues.apache.org/jira/browse/SPARK-10539
I added some comments based on the fix of https://github.com/apache/spark/pull/8742.
Author: Yijie Shen <henry.yijieshen@gmail.com>
Author: Yin Huai <yhuai@databricks.com>
Closes#8823 from yhuai/fix_set_optimization.
Kryo fails with buffer overflow even with max value (2G).
{noformat}
org.apache.spark.SparkException: Kryo serialization failed: Buffer overflow. Available: 0, required: 1
Serialization trace:
containsChild (org.apache.spark.sql.catalyst.expressions.BoundReference)
child (org.apache.spark.sql.catalyst.expressions.SortOrder)
array (scala.collection.mutable.ArraySeq)
ordering (org.apache.spark.sql.catalyst.expressions.InterpretedOrdering)
interpretedOrdering (org.apache.spark.sql.types.StructType)
schema (org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema). To avoid this, increase spark.kryoserializer.buffer.max value.
at org.apache.spark.serializer.KryoSerializerInstance.serialize(KryoSerializer.scala:263)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:240)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
{noformat}
Author: navis.ryu <navis@apache.org>
Closes#8808 from navis/SPARK-10684.
This fixes https://issues.apache.org/jira/browse/SPARK-9794 by using a real ISO8601 parser. (courtesy of the xml component of the standard java library)
cc: angelini
Author: Kevin Cox <kevincox@kevincox.ca>
Closes#8396 from kevincox/kevincox-sql-time-parsing.
Sometimes we can't push down the whole `Project` though `Sort`, but we still have a chance to push down part of it.
Author: Wenchen Fan <cloud0fan@outlook.com>
Closes#8644 from cloud-fan/column-prune.
JIRA: https://issues.apache.org/jira/browse/SPARK-10437
If an expression in `SortOrder` is a resolved one, such as `count(1)`, the corresponding rule in `Analyzer` to make it work in order by will not be applied.
Author: Liang-Chi Hsieh <viirya@appier.com>
Closes#8599 from viirya/orderby-agg.
Move .java files in `src/main/scala` to `src/main/java` root, except for `package-info.java` (to stay next to package.scala)
Author: Sean Owen <sowen@cloudera.com>
Closes#8736 from srowen/SPARK-10576.
Adding STDDEV support for DataFrame using 1-pass online /parallel algorithm to compute variance. Please review the code change.
Author: JihongMa <linlin200605@gmail.com>
Author: Jihong MA <linlin200605@gmail.com>
Author: Jihong MA <jihongma@jihongs-mbp.usca.ibm.com>
Author: Jihong MA <jihongma@Jihongs-MacBook-Pro.local>
Closes#6297 from JihongMA/SPARK-SQL.