| commit | d8c04cf2fb7599c993948df10f4746b70f8c52b9 | [log] [tgz] |
|---|---|---|
| author | Jungtaek Lim <kabhwan.opensource@gmail.com> | Sat Oct 05 07:38:42 2024 +0900 |
| committer | Jungtaek Lim <kabhwan.opensource@gmail.com> | Sat Oct 05 07:38:42 2024 +0900 |
| tree | 9dac87d3be476aece40701e67d0903203d9d3025 | |
| parent | 0c653db0b2fe9d337a1f69220862d05df2f0daf7 [diff] |
[SPARK-49836][SQL][SS] Fix possibly broken query when window is provided to window/session_window fn ### What changes were proposed in this pull request? This PR fixes the correctness issue about losing operators during analysis - it happens when window is provided to window()/session_window() function. The rule `TimeWindowing` and `SessionWindowing` are responsible to resolve the time window functions. When the window function has `window` as parameter (time column) (in other words, building time window from time window), the rule wraps window with WindowTime function so that the rule ResolveWindowTime will further resolve this. (And TimeWindowing/SessionWindowing will resolve this again against the result of ResolveWindowTime.) The issue is that the rule uses "return" for the above, which intends to have "early return" as the other branch is too long compared to this branch. This unfortunately does not work as intended - the intention is just to go out of current local scope (mostly end of curly brace), but it seems to break the loop of execution in "outer" side. (I haven't debugged further but it's simply clear that it doesn't work as intended.) Quoting from Scala doc: > Nonlocal returns are implemented by throwing and catching scala.runtime.NonLocalReturnException-s. It's not super clear where NonLocalReturnException is caught in the call stack; it might exit the execution for much broader scope (context) than expected. And it's finally deprecated in Scala 3.2 and likely be removed in future. https://dotty.epfl.ch/docs/reference/dropped-features/nonlocal-returns.html Interestingly it does not break every query for chained time window aggregations. Spark already has several tests with DataFrame API and they haven't failed. The reproducer in community report is using SQL statement - where each aggregation is considered as subquery. This PR fixes the rule to NOT use early return and instead have a huge if else. ### Why are the changes needed? Described in above. ### Does this PR introduce _any_ user-facing change? Yes, this fixes the possible query breakage. The impacted workloads may not be very huge as chained time window aggregations is an advanced usage, and it does not break every query for the usage. ### How was this patch tested? New UTs. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #48309 from HeartSaVioR/SPARK-49836. Lead-authored-by: Jungtaek Lim <kabhwan.opensource@gmail.com> Co-authored-by: Andrzej Zera <andrzejzera@gmail.com> Signed-off-by: Jungtaek Lim <kabhwan.opensource@gmail.com>
Spark is a unified analytics engine for large-scale data processing. It provides high-level APIs in Scala, Java, Python, and R, and an optimized engine that supports general computation graphs for data analysis. It also supports a rich set of higher-level tools including Spark SQL for SQL and DataFrames, pandas API on Spark for pandas workloads, MLlib for machine learning, GraphX for graph processing, and Structured Streaming for stream processing.
You can find the latest Spark documentation, including a programming guide, on the project web page. This README file only contains basic setup instructions.
Spark is built using Apache Maven. To build Spark and its example programs, run:
./build/mvn -DskipTests clean package
(You do not need to do this if you downloaded a pre-built package.)
More detailed documentation is available from the project site, at “Building Spark”.
For general development tips, including info on developing Spark using an IDE, see “Useful Developer Tools”.
The easiest way to start using Spark is through the Scala shell:
./bin/spark-shell
Try the following command, which should return 1,000,000,000:
scala> spark.range(1000 * 1000 * 1000).count()
Alternatively, if you prefer Python, you can use the Python shell:
./bin/pyspark
And run the following command, which should also return 1,000,000,000:
>>> spark.range(1000 * 1000 * 1000).count()
Spark also comes with several sample programs in the examples directory. To run one of them, use ./bin/run-example <class> [params]. For example:
./bin/run-example SparkPi
will run the Pi example locally.
You can set the MASTER environment variable when running examples to submit examples to a cluster. This can be spark:// URL, “yarn” to run on YARN, and “local” to run locally with one thread, or “local[N]” to run locally with N threads. You can also use an abbreviated class name if the class is in the examples package. For instance:
MASTER=spark://host:7077 ./bin/run-example SparkPi
Many of the example programs print usage help if no params are given.
Testing first requires building Spark. Once Spark is built, tests can be run using:
./dev/run-tests
Please see the guidance on how to run tests for a module, or individual tests.
There is also a Kubernetes integration test, see resource-managers/kubernetes/integration-tests/README.md
Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported storage systems. Because the protocols have changed in different versions of Hadoop, you must build Spark against the same version that your cluster runs.
Please refer to the build documentation at “Specifying the Hadoop Version and Enabling YARN” for detailed guidance on building for a particular distribution of Hadoop, including building for particular Hive and Hive Thriftserver distributions.
Please refer to the Configuration Guide in the online documentation for an overview on how to configure Spark.
Please review the Contribution to Spark guide for information on how to get started contributing to the project.
