ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
spark-instrumented-optimizer/sql
History
Wenchen Fan 8d817dcf30 [SPARK-36315][SQL] Only skip AQEShuffleReadRule in the final stage if it breaks the distribution requirement 
### What changes were proposed in this pull request?

This is a followup of https://github.com/apache/spark/pull/30494

This PR proposes a new way to optimize the final query stage in AQE. We first collect the effective user-specified repartition (semantic-wise, user-specified repartition is only effective if it's the root node or under a few simple nodes), and get the required distribution for the final plan. When we optimize the final query stage, we skip certain `AQEShuffleReadRule` if it breaks the required distribution.

### Why are the changes needed?

The current solution for optimizing the final query stage is pretty hacky and overkill. As an example, the newly added rule `OptimizeSkewInRebalancePartitions` can hardly apply as it's very common that the query plan has shuffles with origin `ENSURE_REQUIREMENTS`, which is not supported by `OptimizeSkewInRebalancePartitions`.

### Does this PR introduce _any_ user-facing change?

no

### How was this patch tested?

updated tests

Closes #33541 from cloud-fan/aqe.

Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit dd80457ffb)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
2021-08-03 18:29:07 +08:00
..
catalyst [SPARK-36380][SQL] Simplify the logical plan names for ALTER TABLE ... COLUMN 2021-08-03 10:43:15 +03:00
core [SPARK-36315][SQL] Only skip AQEShuffleReadRule in the final stage if it breaks the distribution requirement 2021-08-03 18:29:07 +08:00
hive [SPARK-36224][SQL] Use Void as the type name of NullType 2021-08-02 23:20:11 +08:00
hive-thriftserver [SPARK-36224][SQL] Use Void as the type name of NullType 2021-08-02 23:20:11 +08:00
create-docs.sh [SPARK-34010][SQL][DODCS] Use python3 instead of python in SQL documentation build 2021-01-05 19:48:10 +09:00
gen-sql-api-docs.py [SPARK-34747][SQL][DOCS] Add virtual operators to the built-in function document 2021-03-19 10:19:26 +09:00
gen-sql-config-docs.py [SPARK-32194][PYTHON] Use proper exception classes instead of plain Exception 2021-05-26 11:54:40 +09:00
gen-sql-functions-docs.py
mkdocs.yml
README.md

README.md

Spark SQL

This module provides support for executing relational queries expressed in either SQL or the DataFrame/Dataset API.

Spark SQL is broken up into four subprojects:

  • Catalyst (sql/catalyst) - An implementation-agnostic framework for manipulating trees of relational operators and expressions.
  • Execution (sql/core) - A query planner / execution engine for translating Catalyst's logical query plans into Spark RDDs. This component also includes a new public interface, SQLContext, that allows users to execute SQL or LINQ statements against existing RDDs and Parquet files.
  • Hive Support (sql/hive) - Includes extensions that allow users to write queries using a subset of HiveQL and access data from a Hive Metastore using Hive SerDes. There are also wrappers that allow users to run queries that include Hive UDFs, UDAFs, and UDTFs.
  • HiveServer and CLI support (sql/hive-thriftserver) - Includes support for the SQL CLI (bin/spark-sql) and a HiveServer2 (for JDBC/ODBC) compatible server.

Running ./sql/create-docs.sh generates SQL documentation for built-in functions under sql/site, and SQL configuration documentation that gets included as part of configuration.md in the main docs directory.