[SPARK-56638][SQL] Replace legacy parser error for ParserUtils validation failures by XdithyX · Pull Request #55782 · apache/spark

XdithyX · 2026-05-09T15:45:23Z

What changes were proposed in this pull request?

This PR removes the remaining use of _LEGACY_ERROR_TEMP_0064 from parser validation paths that depended on ParserUtils.validate(...) (SPARK-56638).

Specifically, this PR:

Adds named parser error conditions for the affected validation failures
Adds corresponding QueryParsingErrors helpers
Replaces ParserUtils.validate(...) call sites with specific parser errors
Removes ParserUtils.validate(...)
Updates parser tests and SQL golden files to assert the new error classes

This includes the TABLESAMPLE fraction validation path, covering both existing TABLESAMPLE syntax and TABLESAMPLE SYSTEM from #54972.

Why are the changes needed?

ParserUtils.validate(...) always threw _LEGACY_ERROR_TEMP_0064, which made different parser validation failures share the same generic legacy error class.

Using named error classes makes each parser failure explicit and removes this legacy temporary error from the affected parser paths.

INVALID_TABLESAMPLE_FRACTION uses SQLSTATE 22023 because the statement is syntactically valid, but the supplied sampling fraction is invalid/out of range. The other parser syntax validation errors are added under INVALID_SQL_SYNTAX, which uses SQLSTATE 42000.

Does this PR introduce any user-facing change?

Yes. The error class changes from _LEGACY_ERROR_TEMP_0064 to named error classes for the affected parser validation failures.

For invalid TABLESAMPLE fractions, the new error class is INVALID_TABLESAMPLE_FRACTION.

How was this patch tested?

build/sbt 'catalyst / Test / testOnly org.apache.spark.sql.catalyst.parser.PlanParserSuite -- -z "sampled relations"'
build/sbt 'catalyst / Test / testOnly org.apache.spark.sql.catalyst.parser.PlanParserSuite -- -z "TABLESAMPLE SYSTEM - fraction out of range"'

SPARK_GENERATE_GOLDEN_FILES=1 build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "tablesample-negative.sql"'
SPARK_GENERATE_GOLDEN_FILES=1 build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "pipe-operators.sql"'
build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "tablesample-negative.sql"' 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "pipe-operators.sql"'

Was this patch authored or co-authored using generative AI tooling?

Yes. Generated-by: OpenAI GPT-5.5 Codex

XdithyX · 2026-05-09T15:50:20Z

Hi @cloud-fan and @sarutak, could you please take a look? Thanks.
Tagging @HeartSaVioR as well.

stanyao · 2026-05-11T23:24:17Z

@XdithyX , thank you for fixing this issue. Please fix this issue after #54972 merges, because this issue was found during this PR. To keep this already big PR focused and fix it everywhere (including the code that is not part of #54972) I opened this Jira item (SPARK-56638) to track a follow up fix.

XdithyX · 2026-05-12T12:44:14Z

@stanyao Thanks. I rebased this PR after #54972 merged and updated the new TABLESAMPLE SYSTEM out-of-range test to expect INVALID_TABLESAMPLE_FRACTION.
No other changes were required, I also regenerated and verified the affected SQL golden files.

Can you please have a look when you get a chance?

XdithyX · 2026-05-13T20:14:52Z

CI has now passed. Please let me know if anything else is required.
cc: @stanyao

stanyao · 2026-05-15T15:57:53Z

The goal of this JIRA item is to deprecate the usage of "_LEGACY_ERROR_TEMP_0064" in TABLESAMPLE code path and likely other code paths that depend on ParserUtils.scala. For example, in test("SPARK-55978: TABLESAMPLE SYSTEM - fraction out of range") in PlanParserSuite.scala. Starting from that call site, there is a call chain of -> AnalysisTest.scala -> AbstractSqlParser.scala -> AstBuilder.scala -> validate() function in ParserUtils.scala that is shared by both the existing Bernoulli sampling and the new System sampling. Fixing this will need code changes to 10+ call sites and 20+ test cases. That is a super set of TABLESAMPLE code path. Fixing them all at once is the best approach. I don't think your PR currently covers the full fix.

XdithyX · 2026-05-15T17:50:02Z

@stanyao Thanks for clarifying. I originally scoped this PR to the TABLESAMPLE fraction validation path only, but I understand now that the expected fix is to eliminate _LEGACY_ERROR_TEMP_0064 from all current ParserUtils.validate(...) call sites, including TABLESAMPLE SYSTEM/Bernoulli and the other parser validations that share the helper.

I’ll update the PR to cover the full set of validate() call sites and their tests/golden files.

XdithyX · 2026-05-17T16:18:29Z

Hi @stanyao I've updated the PR, can you please have a look when you get a chance?
Thank you

stanyao · 2026-05-19T00:26:30Z

+      ctx)
+  }
+
+  def invalidTableFunctionTableArgumentWithSinglePartitionError(


It's best to have test coverage in PlanParserSuite.scala for the newly introduced error classes here.

@stanyao Thanks for pointing this out. I added PlanParserSuite coverage for the reachable table function parser error path, including the ORDER BY variant.

I also removed the two newly added table-function error classes that were not reachable from SQL parsing because the grammar rejects those forms earlier with PARSE_SYNTAX_ERROR.

stanyao · 2026-05-19T15:55:57Z

Thanks for removing the entire ParserUtils.validate() and making progress of deprecating _LEGACY_ERROR_TEMP_0064! This PR touches the error framework (error-conditions.json, QueryParsingErrors), parser infrastructure (ParserUtils.validate removal, AstBuilder), and a bunch of golden files. Please get feedback from @cloud-fan and @MaxGekk.

XdithyX · 2026-05-19T16:00:18Z

@stanyao Thanks for the review. @cloud-fan @MaxGekk Could you please have a look when you get a chance?
Thank you!

stanyao

Overall looks good. I looped in 2 committers who are experts in this area for more feedback.

XdithyX · 2026-05-24T12:47:38Z

Hi @cloud-fan @MaxGekk Gentle follow up on this PR.

cloud-fan · 2026-05-26T04:12:52Z

cc @srielau

cloud-fan · 2026-05-26T04:13:33Z

BTW PR needs a rebase, ChangelogInfo has been renamed.

…ion validation

XdithyX · 2026-05-26T12:59:07Z

Thanks @cloud-fan. Rebased the PR on latest master to pick up the ChangelogInfo rename.

Also re-ran the parser coverage after rebase:

build/sbt 'catalyst / Test / testOnly org.apache.spark.sql.catalyst.parser.PlanParserSuite -- -z "Support PARTITION BY and ORDER BY clause for TVF TABLE arguments"'

XdithyX · 2026-05-28T17:50:14Z

Gentle follow-up @srielau @cloud-fan . Please let me know if anything else is needed from my side. Thanks!

srielau

LGTM

XdithyX · 2026-05-31T17:56:04Z

Thank you @srielau for the review. @cloud-fan Can we check this in, if there are no other issues?

cloud-fan · 2026-06-01T04:42:52Z

thanks, merging to master/4.x!

…tion failures ### What changes were proposed in this pull request? This PR removes the remaining use of `_LEGACY_ERROR_TEMP_0064` from parser validation paths that depended on `ParserUtils.validate(...)` ([SPARK-56638](https://issues.apache.org/jira/browse/SPARK-56638)). Specifically, this PR: - Adds named parser error conditions for the affected validation failures - Adds corresponding `QueryParsingErrors` helpers - Replaces `ParserUtils.validate(...)` call sites with specific parser errors - Removes `ParserUtils.validate(...)` - Updates parser tests and SQL golden files to assert the new error classes This includes the `TABLESAMPLE` fraction validation path, covering both existing `TABLESAMPLE` syntax and `TABLESAMPLE SYSTEM` from [#54972](#54972). ### Why are the changes needed? `ParserUtils.validate(...)` always threw `_LEGACY_ERROR_TEMP_0064`, which made different parser validation failures share the same generic legacy error class. Using named error classes makes each parser failure explicit and removes this legacy temporary error from the affected parser paths. `INVALID_TABLESAMPLE_FRACTION` uses SQLSTATE `22023` because the statement is syntactically valid, but the supplied sampling fraction is invalid/out of range. The other parser syntax validation errors are added under `INVALID_SQL_SYNTAX`, which uses SQLSTATE `42000`. ### Does this PR introduce _any_ user-facing change? Yes. The error class changes from `_LEGACY_ERROR_TEMP_0064` to named error classes for the affected parser validation failures. For invalid `TABLESAMPLE` fractions, the new error class is `INVALID_TABLESAMPLE_FRACTION`. ### How was this patch tested? ```bash build/sbt 'catalyst / Test / testOnly org.apache.spark.sql.catalyst.parser.PlanParserSuite -- -z "sampled relations"' build/sbt 'catalyst / Test / testOnly org.apache.spark.sql.catalyst.parser.PlanParserSuite -- -z "TABLESAMPLE SYSTEM - fraction out of range"' SPARK_GENERATE_GOLDEN_FILES=1 build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "tablesample-negative.sql"' SPARK_GENERATE_GOLDEN_FILES=1 build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "pipe-operators.sql"' build/sbt 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "tablesample-negative.sql"' 'sql / Test / testOnly org.apache.spark.sql.SQLQueryTestSuite -- -z "pipe-operators.sql"' ``` ### Was this patch authored or co-authored using generative AI tooling? Yes. Generated-by: OpenAI GPT-5.5 Codex Closes #55782 from XdithyX/SPARK-56638. Authored-by: Adithya Ajith <4adhi007@gmail.com> Signed-off-by: Wenchen Fan <wenchen@databricks.com> (cherry picked from commit 447d3b4) Signed-off-by: Wenchen Fan <wenchen@databricks.com>

XdithyX · 2026-06-01T17:45:20Z

Thank you @cloud-fan

zhengruifeng mentioned this pull request May 11, 2026

[SPARK-56943][INFRA] Share SBT precompile artifact with JVM build matrix #55762

Closed

XdithyX force-pushed the SPARK-56638 branch from 01eb7be to cf8966d Compare May 12, 2026 12:32

XdithyX force-pushed the SPARK-56638 branch 2 times, most recently from 99181d7 to 7b22383 Compare May 17, 2026 11:04

XdithyX changed the title ~~[SPARK-56638][SQL] Eliminate legacy error class for TABLESAMPLE fraction validation~~ [SPARK-56638][SQL] Replace legacy parser error for ParserUtils validation failures May 17, 2026

stanyao reviewed May 19, 2026

View reviewed changes

XdithyX requested a review from stanyao May 19, 2026 15:33

stanyao reviewed May 22, 2026

View reviewed changes

XdithyX added 4 commits May 26, 2026 18:23

[SPARK-56638][SQL] Eliminate legacy error class for TABLESAMPLE fract…

4df3fe9

…ion validation

[SPARK-56638][SQL] modify sql.out files of SQLQueryTestSuite

628bb8f

covering full set of validate() call sites

43268f4

adding test coverage in PlanParserSuite

a7de360

XdithyX force-pushed the SPARK-56638 branch from a0b52c0 to a7de360 Compare May 26, 2026 12:57

srielau approved these changes May 31, 2026

View reviewed changes

cloud-fan closed this in 447d3b4 Jun 1, 2026

Conversation

XdithyX commented May 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What changes were proposed in this pull request?

Why are the changes needed?

Does this PR introduce any user-facing change?

How was this patch tested?

Was this patch authored or co-authored using generative AI tooling?

Uh oh!

XdithyX commented May 9, 2026

Uh oh!

stanyao commented May 11, 2026

Uh oh!

XdithyX commented May 12, 2026

Uh oh!

XdithyX commented May 13, 2026

Uh oh!

stanyao commented May 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

XdithyX commented May 15, 2026

Uh oh!

XdithyX commented May 17, 2026

Uh oh!

stanyao May 19, 2026

Choose a reason for hiding this comment

Uh oh!

XdithyX May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

stanyao commented May 19, 2026

Uh oh!

XdithyX commented May 19, 2026

Uh oh!

stanyao left a comment

Choose a reason for hiding this comment

Uh oh!

XdithyX commented May 24, 2026

Uh oh!

cloud-fan commented May 26, 2026

Uh oh!

cloud-fan commented May 26, 2026

Uh oh!

XdithyX commented May 26, 2026

Uh oh!

XdithyX commented May 28, 2026

Uh oh!

srielau left a comment

Choose a reason for hiding this comment

Uh oh!

XdithyX commented May 31, 2026

Uh oh!

cloud-fan commented Jun 1, 2026

Uh oh!

XdithyX commented Jun 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

XdithyX commented May 9, 2026 •

edited

Loading

stanyao commented May 15, 2026 •

edited

Loading

XdithyX May 19, 2026 •

edited

Loading