[SPARK-56929][PYTHON] Pass prefers_large_types when building expected schema for Arrow grouped/cogrouped map UDFs#55961
Closed
Yicong-Huang wants to merge 2 commits into
Closed
Conversation
…/cogrouped map UDFs
| schema = "id long, s string, b binary" | ||
|
|
||
| def func(table): | ||
| assert table.schema.field("s").type == pa.large_string() |
Contributor
Author
There was a problem hiding this comment.
is does not work here as the type from table is not the same instance from the factory function. changed to use pa.types.is_large_binary/is_large_string per offline discussion.
gaogaotiantian
approved these changes
May 18, 2026
zhengruifeng
approved these changes
May 19, 2026
zhengruifeng
pushed a commit
that referenced
this pull request
May 19, 2026
… schema for Arrow grouped/cogrouped map UDFs
### What changes were proposed in this pull request?
Forward `prefers_large_types=runner_conf.use_large_var_types` when building `expected_cols_and_types` in `python/pyspark/worker.py` for `SQL_GROUPED_MAP_ARROW_UDF`, `SQL_GROUPED_MAP_ARROW_ITER_UDF`, and `SQL_COGROUPED_MAP_ARROW_UDF`. The matching `arrow_return_type` already forwards the flag; the per-field expected schema was missing it.
### Why are the changes needed?
With `spark.sql.execution.arrow.useLargeVarTypes=true`, the result table contains `large_string`/`large_binary` (per `arrow_return_type`) while the expected schema contains plain `string`/`binary`, so `verify_arrow_result` raises a spurious `RESULT_COLUMN_TYPES_MISMATCH`:
```python
spark.conf.set("spark.sql.execution.arrow.useLargeVarTypes", True)
df = spark.createDataFrame([(0, "foo", b"foo")], "id long, s string, b binary")
df.groupBy("id").applyInArrow(lambda t: t, "id long, s string, b binary").collect()
# [RESULT_COLUMN_TYPES_MISMATCH] column 's' (expected string, actual large_string), ...
```
Pre-requisite for SPARK-56608.
### Does this PR introduce _any_ user-facing change?
Yes. `applyInArrow` (grouped and cogrouped, iterator and non-iterator) no longer raises a spurious `RESULT_COLUMN_TYPES_MISMATCH` under `useLargeVarTypes=true`. Default behavior unchanged.
### How was this patch tested?
Added `test_apply_in_arrow_large_var_types` to `test_arrow_grouped_map.py` and `test_arrow_cogrouped_map.py`, covering name-based and positional assignment for all three eval types (Spark Connect parity tests pick them up via the mixins). Confirmed the new tests fail on master without the worker.py change and pass with it.
### Was this patch authored or co-authored using generative AI tooling?
No
Closes #55961 from Yicong-Huang/fix-arrow-map-large-var-types.
Authored-by: Yicong Huang <17627829+Yicong-Huang@users.noreply.github.com>
Signed-off-by: Ruifeng Zheng <ruifengz@foxmail.com>
(cherry picked from commit 31fe6dd)
Signed-off-by: Ruifeng Zheng <ruifengz@foxmail.com>
Contributor
|
merged to master/4.x |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Forward
prefers_large_types=runner_conf.use_large_var_typeswhen buildingexpected_cols_and_typesinpython/pyspark/worker.pyforSQL_GROUPED_MAP_ARROW_UDF,SQL_GROUPED_MAP_ARROW_ITER_UDF, andSQL_COGROUPED_MAP_ARROW_UDF. The matchingarrow_return_typealready forwards the flag; the per-field expected schema was missing it.Why are the changes needed?
With
spark.sql.execution.arrow.useLargeVarTypes=true, the result table containslarge_string/large_binary(perarrow_return_type) while the expected schema contains plainstring/binary, soverify_arrow_resultraises a spuriousRESULT_COLUMN_TYPES_MISMATCH:Pre-requisite for SPARK-56608.
Does this PR introduce any user-facing change?
Yes.
applyInArrow(grouped and cogrouped, iterator and non-iterator) no longer raises a spuriousRESULT_COLUMN_TYPES_MISMATCHunderuseLargeVarTypes=true. Default behavior unchanged.How was this patch tested?
Added
test_apply_in_arrow_large_var_typestotest_arrow_grouped_map.pyandtest_arrow_cogrouped_map.py, covering name-based and positional assignment for all three eval types (Spark Connect parity tests pick them up via the mixins). Confirmed the new tests fail on master without the worker.py change and pass with it.Was this patch authored or co-authored using generative AI tooling?
No