### What changes were proposed in this pull request?
Fixing the typos in SQL reference document.
### Why are the changes needed?
For user readability
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Tested manually.
Closes#26424 from jobitmathew/typo.
Authored-by: Jobit Mathew <jobit.mathew@huawei.com>
Signed-off-by: Sean Owen <sean.owen@databricks.com>
### What changes were proposed in this pull request?
Remove the duplicate code
See the build failure: https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Compile/job/spark-master-compile-maven-hadoop-3.2/986/
### Why are the changes needed?
Fix the compilation
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
The existing tests
Closes#26445 from gatorsmile/hotfixPraser.
Authored-by: Xiao Li <gatorsmile@gmail.com>
Signed-off-by: Xiao Li <gatorsmile@gmail.com>
### What changes were proposed in this pull request?
This PR supports MERGE INTO in the parser and add the corresponding logical plan. The SQL syntax likes,
```
MERGE INTO [ds_catalog.][multi_part_namespaces.]target_table [AS target_alias]
USING [ds_catalog.][multi_part_namespaces.]source_table | subquery [AS source_alias]
ON <merge_condition>
[ WHEN MATCHED [ AND <condition> ] THEN <matched_action> ]
[ WHEN MATCHED [ AND <condition> ] THEN <matched_action> ]
[ WHEN NOT MATCHED [ AND <condition> ] THEN <not_matched_action> ]
```
where
```
<matched_action> =
DELETE |
UPDATE SET * |
UPDATE SET column1 = value1 [, column2 = value2 ...]
<not_matched_action> =
INSERT * |
INSERT (column1 [, column2 ...]) VALUES (value1 [, value2 ...])
```
### Why are the changes needed?
This is a start work for introduce `MERGE INTO` support for the builtin datasource, and the design work for the `MERGE INTO` support in DSV2.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
New test cases.
Closes#26167 from xianyinxin/SPARK-28893.
Authored-by: xy_xin <xianyin.xxy@alibaba-inc.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
On the worker we express lambda functions as strings and then eval them to create a "mapper" function. This make the code hard to read & limits the # of arguments a udf can support to 256 for python <= 3.6.
This PR rewrites the mapper functions as nested functions instead of "lambda strings" and allows passing in more than 255 args.
### Why are the changes needed?
The jira ticket associated with this issue describes how MLflow uses udfs to consume columns as features. This pattern isn't unique and a limit of 255 features is quite low.
### Does this PR introduce any user-facing change?
Users can now pass more than 255 cols to a udf function.
### How was this patch tested?
Added a unit test for passing in > 255 args to udf.
Closes#26442 from MrBago/replace-lambdas-on-worker.
Authored-by: Bago Amirbekian <bago@databricks.com>
Signed-off-by: Xiangrui Meng <meng@databricks.com>
### What changes were proposed in this pull request?
Replace qualifiedName with multipartIdentifier in parser rules of DDL commands.
### Why are the changes needed?
There are identifiers in some DDL rules we use `qualifiedName`. We should use `multipartIdentifier` because it can capture wrong identifiers such as `test-table`, `test-col`.
### Does this PR introduce any user-facing change?
Yes. Wrong identifiers such as test-table, will be captured now after this change.
### How was this patch tested?
Unit tests.
Closes#26419 from viirya/SPARK-29680-followup2.
Lead-authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Co-authored-by: Liang-Chi Hsieh <liangchi@uber.com>
Signed-off-by: Liang-Chi Hsieh <liangchi@uber.com>
### What changes were proposed in this pull request?
This PR proposes to infer bytes as binary types in Python 3. See https://github.com/apache/spark/pull/25749 for discussions. I have also checked that Arrow considers `bytes` as binary type, and PySpark UDF can also accepts `bytes` as a binary type.
Since `bytes` is not a `str` anymore in Python 3, it's clear to call it `BinaryType` in Python 3.
### Why are the changes needed?
To respect Python 3's `bytes` type and support Python's primitive types.
### Does this PR introduce any user-facing change?
Yes.
**Before:**
```python
>>> spark.createDataFrame([[b"abc"]])
Traceback (most recent call last):
File "/.../spark/python/pyspark/sql/types.py", line 1036, in _infer_type
return _infer_schema(obj)
File "/.../spark/python/pyspark/sql/types.py", line 1062, in _infer_schema
raise TypeError("Can not infer schema for type: %s" % type(row))
TypeError: Can not infer schema for type: <class 'bytes'>
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/.../spark/python/pyspark/sql/session.py", line 787, in createDataFrame
rdd, schema = self._createFromLocal(map(prepare, data), schema)
File "/.../spark/python/pyspark/sql/session.py", line 445, in _createFromLocal
struct = self._inferSchemaFromList(data, names=schema)
File "/.../spark/python/pyspark/sql/session.py", line 377, in _inferSchemaFromList
schema = reduce(_merge_type, (_infer_schema(row, names) for row in data))
File "/.../spark/python/pyspark/sql/session.py", line 377, in <genexpr>
schema = reduce(_merge_type, (_infer_schema(row, names) for row in data))
File "/.../spark/python/pyspark/sql/types.py", line 1064, in _infer_schema
fields = [StructField(k, _infer_type(v), True) for k, v in items]
File "/.../spark/python/pyspark/sql/types.py", line 1064, in <listcomp>
fields = [StructField(k, _infer_type(v), True) for k, v in items]
File "/.../spark/python/pyspark/sql/types.py", line 1038, in _infer_type
raise TypeError("not supported type: %s" % type(obj))
TypeError: not supported type: <class 'bytes'>
```
**After:**
```python
>>> spark.createDataFrame([[b"abc"]])
DataFrame[_1: binary]
```
### How was this patch tested?
Unittest was added and manually tested.
Closes#26432 from HyukjinKwon/SPARK-29798.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Bryan Cutler <cutlerb@gmail.com>
It adds a note about the required port of a master url in Kubernetes.
Currently a port needs to be specified for the Kubernetes API. Also in case the API is hosted on the HTTPS port. Else the driver might fail with https://medium.com/kidane.weldemariam_75349/thanks-james-on-issuing-spark-submit-i-run-into-this-error-cc507d4f8f0d
Yes, a change to the "Running on Kubernetes" guide.
None - Documentation change
Closes#26426 from Tapped/patch-1.
Authored-by: Emil Sandstø <emilalexer@hotmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
interval type support >, >=, <, <=, =, <=>, order by, min,max..
### Why are the changes needed?
Part of SPARK-27764 Feature Parity between PostgreSQL and Spark
### Does this PR introduce any user-facing change?
yes, we now support compare intervals
### How was this patch tested?
add ut
Closes#26337 from yaooqinn/SPARK-29679.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
avg aggregate support interval type values
### Why are the changes needed?
Part of SPARK-27764 Feature Parity between PostgreSQL and Spark
### Does this PR introduce any user-facing change?
yes, we can do avg on intervals
### How was this patch tested?
add ut
Closes#26347 from yaooqinn/SPARK-29688.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Spark org.apache.spark.sql.functions do not have `if` function so conditions are expressed using `when-otherwise` function. However `If` (which is available in SQL) has more efficient code gen. This pr rewrites `when-otherwise` conditions to `If` if it is possible (`when-otherwise` with single branch)
### Why are the changes needed?
It is an optimization enhancement. Here is a simple performance comparison (tested in local mode (with 4 cores)):
```
val df = spark.range(10000000000L).withColumn("x", rand)
val resultA = df.withColumn("r", when($"x" < 0.5, lit(1)).otherwise(lit(0))).agg(sum($"r"))
val resultB = df.withColumn("r", expr("if(x < 0.5, 1, 0)")).agg(sum($"r"))
resultA.collect() // takes 56s to finish
resultB.collect() // takes 30s to finish
```
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
New test is added.
Closes#26294 from davidvrba/spark-28477_rewriteCaseWhenToIf.
Authored-by: davidvrba <vrba.dave@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1,unpersist intermediate rdd `wordCounts`
2,if the `dataset` is already persisted, we do not need to persist rdd `input`
3,if both `minDF`&`maxDF` are gteq or lt than 1, we can compare & check them af first.
### Why are the changes needed?
we should unpersit unused rdd ASAP
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
existing testsuites
Closes#26398 from zhengruifeng/CountVectorizer_unpersist_wordCounts.
Authored-by: zhengruifeng <ruifengz@foxmail.com>
Signed-off-by: zhengruifeng <ruifengz@foxmail.com>
### What changes were proposed in this pull request?
V2 catalog support namespace, we should add `withNamespace` like `withDatabase`.
### Why are the changes needed?
Make test easy.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Add UT.
Closes#26411 from ulysses-you/Add-test-with-namespace.
Authored-by: ulysses <youxiduo@weidian.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Move method add/subtract/negate from CalendarInterval to IntervalUtils
### Why are the changes needed?
https://github.com/apache/spark/pull/26410#discussion_r343125468 suggested here
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
add uts and move some
Closes#26423 from yaooqinn/SPARK-29787.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Kafka producers are now closed when `spark.kafka.producer.cache.timeout` reached which could be significant problem when processing big SQL queries. The workaround was to increase `spark.kafka.producer.cache.timeout` to a number where the biggest SQL query can be finished.
In this PR I've adapted similar solution which already exists on the consumer side, namely applies Apache Commons Pool on the producer side as well. Main advantages choosing this solution:
* Producers are not closed until they're in use
* No manual reference counting needed (which may be error prone)
* Thread-safe by design
* Provides jmx connection to the pool where metrics can be fetched
What this PR contains:
* Introduced producer side parameters to configure pool
* Renamed `InternalKafkaConsumerPool` to `InternalKafkaConnectorPool` and made it abstract
* Created 2 implementations from it: `InternalKafkaConsumerPool` and `InternalKafkaProducerPool`
* Adapted `CachedKafkaProducer` to use `InternalKafkaProducerPool`
* Changed `KafkaDataWriter` and `KafkaDataWriteTask` to release producer even in failure scenario
* Added several new tests
* Extended `KafkaTest` to clear not only producers but consumers as well
* Renamed `InternalKafkaConsumerPoolSuite` to `InternalKafkaConnectorPoolSuite` where only consumer tests are checking the behavior (please see comment for reasoning)
What this PR not yet contains(but intended when the main concept is stable):
* User facing documentation
### Why are the changes needed?
Kafka producer closed after 10 minutes (with default settings).
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing + additional unit tests.
Cluster tests being started.
Closes#25853 from gaborgsomogyi/SPARK-21869.
Authored-by: Gabor Somogyi <gabor.g.somogyi@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
## What changes were proposed in this pull request?
This PR proposes to add **Single threading model design (pinned thread model)** mode which is an experimental mode to sync threads on PVM and JVM. See https://www.py4j.org/advanced_topics.html#using-single-threading-model-pinned-thread
### Multi threading model
Currently, PySpark uses this model. Threads on PVM and JVM are independent. For instance, in a different Python thread, callbacks are received and relevant Python codes are executed. JVM threads are reused when possible.
Py4J will create a new thread every time a command is received and there is no thread available. See the current model we're using - https://www.py4j.org/advanced_topics.html#the-multi-threading-model
One problem in this model is that we can't sync threads on PVM and JVM out of the box. This leads to some problems in particular at some codes related to threading in JVM side. See:
7056e004ee/core/src/main/scala/org/apache/spark/SparkContext.scala (L334)
Due to reusing JVM threads, seems the job groups in Python threads cannot be set in each thread as described in the JIRA.
### Single threading model design (pinned thread model)
This mode pins and syncs the threads on PVM and JVM to work around the problem above. For instance, in the same Python thread, callbacks are received and relevant Python codes are executed. See https://www.py4j.org/advanced_topics.html#the-single-threading-model
Even though this mode can sync threads on PVM and JVM for other thread related code paths,
this might cause another problem: seems unable to inherit properties as below (assuming multi-thread mode still creates new threads when existing threads are busy, I suspect this issue already exists when multiple jobs are submitted in multi-thread mode; however, it can be always seen in single threading mode):
```bash
$ PYSPARK_PIN_THREAD=true ./bin/pyspark
```
```python
import threading
spark.sparkContext.setLocalProperty("a", "hi")
def print_prop():
print(spark.sparkContext.getLocalProperty("a"))
threading.Thread(target=print_prop).start()
```
```
None
```
Unlike Scala side:
```scala
spark.sparkContext.setLocalProperty("a", "hi")
new Thread(new Runnable {
def run() = println(spark.sparkContext.getLocalProperty("a"))
}).start()
```
```
hi
```
This behaviour potentially could cause weird issues but this PR currently does not target this fix this for now since this mode is experimental.
### How does this PR fix?
Basically there are two types of Py4J servers `GatewayServer` and `ClientServer`. The former is for multi threading and the latter is for single threading. This PR adds a switch to use the latter.
In Scala side:
The logic to select a server is encapsulated in `Py4JServer` and use `Py4JServer` at `PythonRunner` for Spark summit and `PythonGatewayServer` for Spark shell. Each uses `ClientServer` when `PYSPARK_PIN_THREAD` is `true` and `GatewayServer` otherwise.
In Python side:
Simply do an if-else to switch the server to talk. It uses `ClientServer` when `PYSPARK_PIN_THREAD` is `true` and `GatewayServer` otherwise.
This is disabled by default for now.
## How was this patch tested?
Manually tested. This can be tested via:
```python
PYSPARK_PIN_THREAD=true ./bin/pyspark
```
and/or
```bash
cd python
./run-tests --python-executables=python --testnames "pyspark.tests.test_pin_thread"
```
Also, ran the Jenkins tests with `PYSPARK_PIN_THREAD` enabled.
Closes#24898 from HyukjinKwon/pinned-thread.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This aims to exclude the `preview` release to recover `HiveExternalCatalogVersionsSuite`. Currently, new preview release breaks `branch-2.4` PRBuilder since yesterday. New release (especially `preview`) should not affect `branch-2.4`.
- https://github.com/apache/spark/pull/26417 (Failed 4 times)
### Why are the changes needed?
**BEFORE**
```scala
scala> scala.io.Source.fromURL("https://dist.apache.org/repos/dist/release/spark/").mkString.split("\n").filter(_.contains("""<li><a href="spark-""")).map("""<a href="spark-(\d.\d.\d)/">""".r.findFirstMatchIn(_).get.group(1))
java.util.NoSuchElementException: None.get
```
**AFTER**
```scala
scala> scala.io.Source.fromURL("https://dist.apache.org/repos/dist/release/spark/").mkString.split("\n").filter(_.contains("""<li><a href="spark-""")).filterNot(_.contains("preview")).map("""<a href="spark-(\d.\d.\d)/">""".r.findFirstMatchIn(_).get.group(1))
res5: Array[String] = Array(2.3.4, 2.4.4)
```
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
This should pass the PRBuilder.
Closes#26428 from dongjoon-hyun/SPARK-HiveExternalCatalogVersionsSuite.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
```java
public static final int YEARS_PER_DECADE = 10;
public static final int YEARS_PER_CENTURY = 100;
public static final int YEARS_PER_MILLENNIUM = 1000;
public static final byte MONTHS_PER_QUARTER = 3;
public static final int MONTHS_PER_YEAR = 12;
public static final byte DAYS_PER_WEEK = 7;
public static final long DAYS_PER_MONTH = 30L;
public static final long HOURS_PER_DAY = 24L;
public static final long MINUTES_PER_HOUR = 60L;
public static final long SECONDS_PER_MINUTE = 60L;
public static final long SECONDS_PER_HOUR = MINUTES_PER_HOUR * SECONDS_PER_MINUTE;
public static final long SECONDS_PER_DAY = HOURS_PER_DAY * SECONDS_PER_HOUR;
public static final long MILLIS_PER_SECOND = 1000L;
public static final long MILLIS_PER_MINUTE = SECONDS_PER_MINUTE * MILLIS_PER_SECOND;
public static final long MILLIS_PER_HOUR = MINUTES_PER_HOUR * MILLIS_PER_MINUTE;
public static final long MILLIS_PER_DAY = HOURS_PER_DAY * MILLIS_PER_HOUR;
public static final long MICROS_PER_MILLIS = 1000L;
public static final long MICROS_PER_SECOND = MILLIS_PER_SECOND * MICROS_PER_MILLIS;
public static final long MICROS_PER_MINUTE = SECONDS_PER_MINUTE * MICROS_PER_SECOND;
public static final long MICROS_PER_HOUR = MINUTES_PER_HOUR * MICROS_PER_MINUTE;
public static final long MICROS_PER_DAY = HOURS_PER_DAY * MICROS_PER_HOUR;
public static final long MICROS_PER_MONTH = DAYS_PER_MONTH * MICROS_PER_DAY;
/* 365.25 days per year assumes leap year every four years */
public static final long MICROS_PER_YEAR = (36525L * MICROS_PER_DAY) / 100;
public static final long NANOS_PER_MICROS = 1000L;
public static final long NANOS_PER_MILLIS = MICROS_PER_MILLIS * NANOS_PER_MICROS;
public static final long NANOS_PER_SECOND = MILLIS_PER_SECOND * NANOS_PER_MILLIS;
```
The above parameters are defined in IntervalUtils, DateTimeUtils, and CalendarInterval, some of them are redundant, some of them are cross-referenced.
### Why are the changes needed?
To simplify code, enhance consistency and reduce risks
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
modified uts
Closes#26399 from yaooqinn/SPARK-29757.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
remove the leading "interval" in `CalendarInterval.toString`.
### Why are the changes needed?
Although it's allowed to have "interval" prefix when casting string to int, it's not recommended.
This is also consistent with pgsql:
```
cloud0fan=# select interval '1' day;
interval
----------
1 day
(1 row)
```
### Does this PR introduce any user-facing change?
yes, when display a dataframe with interval type column, the result is different.
### How was this patch tested?
updated tests.
Closes#26401 from cloud-fan/interval.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose new function `stringToInterval()` in `IntervalUtils` for converting `UTF8String` to `CalendarInterval`. The function is used in casting a `STRING` column to an `INTERVAL` column.
### Why are the changes needed?
The proposed implementation is ~10 times faster. For example, parsing 9 interval units on JDK 8:
Before:
```
9 units w/ interval 14004 14125 116 0.1 14003.6 0.0X
9 units w/o interval 13785 14056 290 0.1 13784.9 0.0X
```
After:
```
9 units w/ interval 1343 1344 1 0.7 1343.0 0.3X
9 units w/o interval 1345 1349 8 0.7 1344.6 0.3X
```
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
- By new tests for `stringToInterval` in `IntervalUtilsSuite`
- By existing tests
Closes#26256 from MaxGekk/string-to-interval.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Handle the inconsistence dividing zeros between literals and columns.
fix the null issue too.
### Why are the changes needed?
BUG FIX
### 1 Handle the inconsistence dividing zeros between literals and columns
```sql
-- !query 24
select
k,
v,
cast(k as interval) / v,
cast(k as interval) * v
from VALUES
('1 seconds', 1),
('2 seconds', 0),
('3 seconds', null),
(null, null),
(null, 0) t(k, v)
-- !query 24 schema
struct<k:string,v:int,divide_interval(CAST(k AS INTERVAL), CAST(v AS DOUBLE)):interval,multiply_interval(CAST(k AS INTERVAL), CAST(v AS DOUBLE)):interval>
-- !query 24 output
1 seconds 1 interval 1 seconds interval 1 seconds
2 seconds 0 interval 0 microseconds interval 0 microseconds
3 seconds NULL NULL NULL
NULL 0 NULL NULL
NULL NULL NULL NULL
```
```sql
-- !query 21
select interval '1 year 2 month' / 0
-- !query 21 schema
struct<divide_interval(interval 1 years 2 months, CAST(0 AS DOUBLE)):interval>
-- !query 21 output
NULL
```
in the first case, interval ’2 seconds ‘ / 0, it produces `interval 0 microseconds `
in the second case, it is `null`
### 2 null literal issues
```sql
-- !query 20
select interval '1 year 2 month' / null
-- !query 20 schema
struct<>
-- !query 20 output
org.apache.spark.sql.AnalysisException
cannot resolve '(interval 1 years 2 months / NULL)' due to data type mismatch: differing types in '(interval 1 years 2 months / NULL)' (interval and null).; line 1 pos 7
-- !query 22
select interval '4 months 2 weeks 6 days' * null
-- !query 22 schema
struct<>
-- !query 22 output
org.apache.spark.sql.AnalysisException
cannot resolve '(interval 4 months 20 days * NULL)' due to data type mismatch: differing types in '(interval 4 months 20 days * NULL)' (interval and null).; line 1 pos 7
-- !query 23
select null * interval '4 months 2 weeks 6 days'
-- !query 23 schema
struct<>
-- !query 23 output
org.apache.spark.sql.AnalysisException
cannot resolve '(NULL * interval 4 months 20 days)' due to data type mismatch: differing types in '(NULL * interval 4 months 20 days)' (null and interval).; line 1 pos 7
```
dividing or multiplying null literals, error occurs; where in column is fine as the first case
### Does this PR introduce any user-facing change?
NO, maybe yes, but it is just a follow-up
### How was this patch tested?
add uts
cc cloud-fan MaxGekk maropu
Closes#26410 from yaooqinn/SPARK-29387.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This patch leverages V2 continuous memory stream to extract tests from Kafka micro-batch sink suite and continuous sink suite and deduplicate them. These tests are basically doing the same, except how to run and verify the result.
### Why are the changes needed?
We no longer have same tests spotted on two places - brings 300 lines deletion.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing UTs.
Closes#26292 from HeartSaVioR/SPARK-29635.
Authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
Update `LocalShuffleReaderExec.outputPartitioning` to use attributes from `ReusedQueryStage`.
This also removes the override `doCanonicalize` in local/coalesced shuffle reader, as these 2 operators change the output partitioning. It's not safe to strip them in the canonicalized query plan.
### Why are the changes needed?
We will have an invalid output partitioning if we don fix it.
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
existing tests
Closes#26400 from cloud-fan/aqe.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Xiao Li <gatorsmile@gmail.com>
### What changes were proposed in this pull request?
This patch fixes the bug that `ContinuousMemoryStream[String]` throws error regarding ClassCastException - cast String to UTFString. This is because ContinuousMemoryStream and ContinuousRecordEndpoint uses origin input as it is for underlying data structure of Row, and encoding is missing here.
To force encoding, this patch changes the element type of underlying array to UnsafeRow instead of Any for ContinuousRecordEndpoint - ContinuousMemoryStream and TextSocketContinuousStream are modified to reflect the change.
### Why are the changes needed?
Above section describes the bug.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Add new UT to check for availability on couple of types.
Closes#26300 from HeartSaVioR/SPARK-29642.
Authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
instead of checking the exact number of local shuffle readers, we should check whether the number of shuffles is equal to the number of local readers.
### Why are the changes needed?
AQE is known to have randomness. We may pick different build side for broadcast join depending on which query stage finishes first. The decision to build side may add/remove shuffles downstream, so it's flaky to check the exact number of local shuffle readers.
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
test only PR.
Closes#26394 from cloud-fan/test.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Xiao Li <gatorsmile@gmail.com>
### What changes were proposed in this pull request?
Priority for YARN to define pending applications ordering policy, those with higher priority have a better opportunity to be activated. YARN CapacityScheduler only.
### Why are the changes needed?
Ordering pending spark apps
### Does this PR introduce any user-facing change?
add a conf
### How was this patch tested?
add ut
Closes#26255 from yaooqinn/SPARK-29603.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
This PR change the Github Actions build command from `mvn package` to `mvn install` to build Scaladoc jars.
### Why are the changes needed?
Sometimes `mvn install` build failure with error: `not found: type ClassName...`.
More details: https://github.com/apache/spark/pull/24628#issuecomment-495655747
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
N/A
Closes#26414 from wangyum/github-action-install.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
use `ml.Summarizer` instead of `mllib.MultivariateOnlineSummarizer`
### Why are the changes needed?
1, I found that using `ml.Summarizer` is faster than current impl;
2, `mllib.MultivariateOnlineSummarizer` maintain all arrays, while `ml.Summarizer` only maintain necessary arrays
3, using `ml.Summarizer` will avoid vector conversions to `mlllib.Vector`
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
existing testsuites
Closes#26393 from zhengruifeng/maxabs_opt.
Authored-by: zhengruifeng <ruifengz@foxmail.com>
Signed-off-by: Sean Owen <sean.owen@databricks.com>
### What changes were proposed in this pull request?
This PR implements ```validateInput``` in ```ElementwiseProduct```, ```Normalizer``` and ```PolynomialExpansion```.
### Why are the changes needed?
```UnaryTransformer``` has abstract method ```validateInputType``` and call it in ```transformSchema```, but this method is not implemented in ```ElementwiseProduct```, ```Normalizer``` and ```PolynomialExpansion```.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Existing tests
Closes#26388 from huaxingao/spark-29746.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: Sean Owen <sean.owen@databricks.com>
### What changes were proposed in this pull request?
Added UT for the classes `ThriftServerPage.scala` and `ThriftServerSessionPage.scala`
### Why are the changes needed?
Currently, there are no UTs for testing Thriftserver UI page
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
UT
Closes#26403 from shahidki31/ut.
Authored-by: shahid <shahidki31@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
1, change the scope of `ml.SummarizerBuffer` and add a method `createSummarizerBuffer` for it, so it can be used as an aggregator like `MultivariateOnlineSummarizer`;
2, In LoR/AFT/LiR/SVC, use Summarizer instead of MultivariateOnlineSummarizer
### Why are the changes needed?
The computation of summary before learning iterations is a bottleneck in high-dimension cases, since `MultivariateOnlineSummarizer` compute much more than needed.
In the [ticket](https://issues.apache.org/jira/browse/SPARK-29754) is an example, with `--driver-memory=4G` LoR will always fail on KDDA dataset. If we swith to `ml.Summarizer`, then `--driver-memory=3G` is enough to train a model.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
existing testsuites & manual test in REPL
Closes#26396 from zhengruifeng/using_SummarizerBuffer.
Authored-by: zhengruifeng <ruifengz@foxmail.com>
Signed-off-by: zhengruifeng <ruifengz@foxmail.com>
### What changes were proposed in this pull request?
During creation of array, if CreateArray does not gets any children to set data type for array, it will create an array of null type .
### Why are the changes needed?
When empty array is created, it should be declared as array<null>.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Tested manually
Closes#26324 from amanomer/29462.
Authored-by: Aman Omer <amanomer1996@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This patch removes v1 ALTER TABLE CHANGE COLUMN syntax.
### Why are the changes needed?
Since in v2 we have ALTER TABLE CHANGE COLUMN and ALTER TABLE RENAME COLUMN, this old syntax is not necessary now and can be confusing.
The v2 ALTER TABLE CHANGE COLUMN should fallback to v1 AlterTableChangeColumnCommand (#26354).
### Does this PR introduce any user-facing change?
Yes, the old v1 ALTER TABLE CHANGE COLUMN syntax is removed.
### How was this patch tested?
Unit tests.
Closes#26338 from viirya/SPARK-29680.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
expose expert param `aggregationDepth` in algs: GMM/GLR
### Why are the changes needed?
SVC/LoR/LiR/AFT had exposed expert param aggregationDepth to end users. It should be nice to expose it in similar algs.
### Does this PR introduce any user-facing change?
yes, expose new param
### How was this patch tested?
added pytext tests
Closes#26322 from zhengruifeng/agg_opt.
Authored-by: zhengruifeng <ruifengz@foxmail.com>
Signed-off-by: zhengruifeng <ruifengz@foxmail.com>
### What changes were proposed in this pull request?
Fix the task table UI to show all accumulators.
Below example was creating 2 accumulators
scala> val accum = sc.longAccumulator("My Accumulator")
scala> val accum2 = sc.longAccumulator("My Accumulator")
scala> sc.parallelize(Array(1, 2, 3, 4)).foreach(x => {
accum2.add(x)
accum.add(x)
})
Before this change, only shows a single on in task table:
After this change you can see all of them:
### Why are the changes needed?
Its not showing all accumulators now.
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
Manual testing the UI.
Closes#26402 from tgravescs/SPARK-29763.
Authored-by: Thomas Graves <tgraves@nvidia.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
Added new expressions `MultiplyInterval` and `DivideInterval` to multiply/divide an interval by a numeric. Updated `TypeCoercion.DateTimeOperations` to turn the `Multiply`/`Divide` expressions of `CalendarIntervalType` and `NumericType` to `MultiplyInterval`/`DivideInterval`.
To support new operations, added new methods `multiply()` and `divide()` to `CalendarInterval`.
### Why are the changes needed?
- To maintain feature parity with PostgreSQL which supports multiplication and division of intervals by doubles:
```sql
# select interval '1 hour' / double precision '1.5';
?column?
----------
00:40:00
```
- To conform the SQL standard which defines those operations: `numeric * interval`, `interval * numeric` and `interval / numeric`. See [4.5.3 Operations involving datetimes and intervals](http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt).
- Improve Spark SQL UX and allow users to adjust interval columns. For example:
```sql
spark-sql> select (timestamp'now' - timestamp'yesterday') * 1.3;
interval 2 days 10 hours 39 minutes 38 seconds 568 milliseconds 900 microseconds
```
### Does this PR introduce any user-facing change?
Yes, previously the following query fails with the error:
```sql
spark-sql> select interval 1 hour 30 minutes * 1.5;
Error in query: cannot resolve '(interval 1 hours 30 minutes * 1.5BD)' due to data type mismatch: differing types in '(interval 1 hours 30 minutes * 1.5BD)' (interval and decimal(2,1)).; line 1 pos 7;
```
After:
```sql
spark-sql> select interval 1 hour 30 minutes * 1.5;
interval 2 hours 15 minutes
```
### How was this patch tested?
- Added tests for the `multiply()` and `divide()` methods to `CalendarIntervalSuite.java`
- New test suite `IntervalExpressionsSuite`
- by tests for `Multiply` -> `MultiplyInterval` and `Divide` -> `DivideInterval` in `TypeCoercionSuite`
- updated `datetime.sql`
Closes#26132 from MaxGekk/interval-mul-div.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR adds the ability for tasks to request fractional resources, in order to be able to execute more than 1 task per resource. For example, if you have 1 GPU in the executor, and the task configuration is 0.5 GPU/task, the executor can schedule two tasks to run on that 1 GPU.
### Why are the changes needed?
Currently there is no good way to share a resource such that multiple tasks can run on a single unit. This allows multiple tasks to share an executor resource.
### Does this PR introduce any user-facing change?
Yes: There is a configuration change where `spark.task.resource.[resource type].amount` can now be fractional.
### How was this patch tested?
Unit tests and manually on standalone mode, and yarn.
Closes#26078 from abellina/SPARK-29151.
Authored-by: Alessandro Bellina <abellina@nvidia.com>
Signed-off-by: Thomas Graves <tgraves@apache.org>
### What changes were proposed in this pull request?
Add AlterTableSerDePropertiesStatement and make ALTER TABLE ... SET SERDE/SERDEPROPERTIES go through the same catalog/table resolution framework of v2 commands.
### Why are the changes needed?
It's important to make all the commands have the same table resolution behavior, to avoid confusing end-users. e.g.
```
USE my_catalog
DESC t // success and describe the table t from my_catalog
ALTER TABLE t SET SERDE 'org.apache.class' // report table not found as there is no table t in the session catalog
```
### Does this PR introduce any user-facing change?
Yes. When running ALTER TABLE ... SET SERDE/SERDEPROPERTIES, Spark fails the command if the current catalog is set to a v2 catalog, or the table name specified a v2 catalog.
### How was this patch tested?
Unit tests.
Closes#26374 from huaxingao/spark_29695.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Disallow creating a permanent view that references a temporary view in **expressions**.
### Why are the changes needed?
Creating a permanent view that references a temporary view is currently disallowed. For example,
```SQL
# The following throws org.apache.spark.sql.AnalysisException
# Not allowed to create a permanent view `per_view` by referencing a temporary view `tmp`;
CREATE VIEW per_view AS SELECT t1.a, t2.b FROM base_table t1, (SELECT * FROM tmp) t2"
```
However, the following is allowed.
```SQL
CREATE VIEW per_view AS SELECT * FROM base_table WHERE EXISTS (SELECT * FROM tmp);
```
This PR fixes the bug where temporary views used inside expressions are not checked.
### Does this PR introduce any user-facing change?
Yes. Now the following SQL query throws an exception as expected:
```SQL
# The following throws org.apache.spark.sql.AnalysisException
# Not allowed to create a permanent view `per_view` by referencing a temporary view `tmp`;
CREATE VIEW per_view AS SELECT * FROM base_table WHERE EXISTS (SELECT * FROM tmp);
```
### How was this patch tested?
Added new unit tests.
Closes#26361 from imback82/spark-29630.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR introduces a new SQL command: `SHOW CURRENT NAMESPACE`.
### Why are the changes needed?
Datasource V2 supports multiple catalogs/namespaces and having `SHOW CURRENT NAMESPACE` to retrieve the current catalog/namespace info would be useful.
### Does this PR introduce any user-facing change?
Yes, the user can perform the following:
```
scala> spark.sql("SHOW CURRENT NAMESPACE").show
+-------------+---------+
| catalog|namespace|
+-------------+---------+
|spark_catalog| default|
+-------------+---------+
scala> spark.sql("USE testcat.ns1.ns2").show
scala> spark.sql("SHOW CURRENT NAMESPACE").show
+-------+---------+
|catalog|namespace|
+-------+---------+
|testcat| ns1.ns2|
+-------+---------+
```
### How was this patch tested?
Added unit tests.
Closes#26379 from imback82/show_current_catalog.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This upgrades joda-time from 2.9 to 2.10.5.
### Why are the changes needed?
Joda 2.9 is almost 4 yrs ago and there are bugs fix and tz database updates.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Existing tests.
Closes#26389 from viirya/upgrade-joda.
Authored-by: Liang-Chi Hsieh <liangchi@uber.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
## What changes were proposed in this pull request?
This patch adds the option to clean up files which are completed in previous batch.
`cleanSource` -> "archive" / "delete" / "off"
The default value is "off", which Spark will do nothing.
If "delete" is specified, Spark will simply delete input files. If "archive" is specified, Spark will require additional config `sourceArchiveDir` which will be used to move input files to there. When archiving (via move) the path of input files are retained to the archived paths as sub-path.
Note that it is only applied to "micro-batch", since for batch all input files must be kept to get same result across multiple query executions.
## How was this patch tested?
Added UT. Manual test against local disk as well as HDFS.
Closes#22952 from HeartSaVioR/SPARK-20568.
Lead-authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
Co-authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan@gmail.com>
Co-authored-by: Jungtaek Lim <kabhwan@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
## What changes were proposed in this pull request?
This pr proposes to be case insensitive when matching dialects via jdbc url prefix.
When I use jdbc url such as: ```jdbc: MySQL://localhost/db``` to query data through sparksql, the result is wrong, but MySQL supports such url writing.
because sparksql matches MySQLDialect by prefix ```jdbc:mysql```, so ```jdbc: MySQL``` is not matched with the correct dialect. Therefore, it should be case insensitive when identifying the corresponding dialect through jdbc url
https://issues.apache.org/jira/browse/SPARK-28552
## How was this patch tested?
UT.
Closes#25287 from teeyog/sql_dialect.
Lead-authored-by: yong.tian1 <yong.tian1@dmall.com>
Co-authored-by: Xingbo Jiang <xingbo.jiang@databricks.com>
Co-authored-by: Chris Martin <chris@cmartinit.co.uk>
Co-authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Co-authored-by: Dongjoon Hyun <dhyun@apple.com>
Co-authored-by: Kent Yao <yaooqinn@hotmail.com>
Co-authored-by: teeyog <teeyog@gmail.com>
Co-authored-by: Maxim Gekk <max.gekk@gmail.com>
Co-authored-by: Ryan Blue <blue@apache.org>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
Spark 2.4 added the ability for executor plugins to be loaded into
Spark (see SPARK-24918). That feature intentionally skipped the
driver to keep changes small, and also because it is possible to
load code into the Spark driver using listeners + configuration.
But that is a bit awkward, because the listener interface does not
provide hooks into a lot of Spark functionality. This change reworks
the executor plugin interface to also extend to the driver.
- there's a "SparkPlugin" main interface that provides APIs to
load driver and executor components.
- custom metric support (added in SPARK-28091) can be used by
plugins to register metrics both in the driver process and in
executors.
- a communication channel now exists that allows the plugin's
executor components to send messages to the plugin's driver
component easily, using the existing Spark RPC system.
The latter was a feature intentionally left out of the original
plugin design (also because it didn't include a driver component).
To avoid polluting the "org.apache.spark" namespace, I added the new
interfaces to the "org.apache.spark.api" package, which seems like
a better place in any case. The actual implementation is kept in
an internal package.
The change includes unit tests for the new interface and features,
but I've also been running a custom plugin that extends the new
API in real applications.
Closes#26170 from vanzin/SPARK-29397.
Authored-by: Marcelo Vanzin <vanzin@cloudera.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
I propose 2 new methods for `CalendarInterval`:
- `extractAsPeriod()` returns the date part of an interval as an instance of `java.time.Period`
- `extractAsDuration()` returns the time part of an interval as an instance of `java.time.Duration`
For example:
```scala
scala> import org.apache.spark.unsafe.types.CalendarInterval
scala> import java.time._
scala> val i = spark.sql("select interval 1 year 3 months 4 days 10 hours 30 seconds").collect()(0).getAs[CalendarInterval](0)
scala> LocalDate.of(2019, 11, 1).plus(i.period())
res8: java.time.LocalDate = 2021-02-05
scala> ZonedDateTime.parse("2019-11-01T12:13:14Z").plus(i.extractAsPeriod()).plus(i.extractAsDuration())
res9: java.time.ZonedDateTime = 2021-02-05T22:13:44Z
```
### Why are the changes needed?
Taking into account that `CalendarInterval` has been already partially exposed to users via the collect operation, and probably it will be fully exposed in the future, it could be convenient for users to get the date and time parts of intervals as java classes:
- to avoid unnecessary dependency from Spark's classes in user code
- to easily use external libraries that accept standard Java classes.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By new test in `CalendarIntervalSuite`.
Closes#26368 from MaxGekk/interval-java-period-duration.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
`SampleExec` has a bug that it sets `needCopyResult` to false as long as the `withReplacement` parameter is false. This causes problems if its child needs to copy the result, e.g. a join.
### Why are the changes needed?
to fix a correctness issue
### Does this PR introduce any user-facing change?
Yes, the result will be corrected.
### How was this patch tested?
a new test
Closes#26387 from cloud-fan/sample-bug.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
