Fix lack of support for new TimestampNTZType in Spark 3.4 datatypes #1385
Conversation
This file contains 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
Signed-off-by: Filipe Oliveira <[email protected]>
Codecov ReportAll modified lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1385 +/- ##
=======================================
Coverage 93.92% 93.92%
=======================================
Files 91 91
Lines 6781 6787 +6
=======================================
+ Hits 6369 6375 +6
Misses 412 412
☔ View full report in Codecov by Sentry. |
…mestampntztype Signed-off-by: Filipe Oliveira <[email protected]>
Signed-off-by: Filipe Oliveira <[email protected]>
cosmicBboy
approved these changes
Oct 19, 2023
|
Thanks, and congrats on your first PR @filipeo2-mck ! 🚀 |
|
Amazing stuff @filipeo2-mck. Thanks for the effort! |
noklam
pushed a commit
to noklam/pandera
that referenced
this pull request
Oct 29, 2023
…nionai-oss#1385) * add TimestampNTZType as equivalents and add parameters to test case Signed-off-by: Filipe Oliveira <[email protected]> * parse version to improve robustness Signed-off-by: Filipe Oliveira <[email protected]> --------- Signed-off-by: Filipe Oliveira <[email protected]> Signed-off-by: Nok <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
When loading a parquet file with timestamp fields into a dataframe using Spark 3.4, pyspark can make use of the new
pyspark.sql.types.TimestampNTZType()but these news fields were not declared as equivalent to Pandera's Timestamp.Stack Trace
Data type '<class 'pyspark.sql.types.TimestampNTZType'>' not understood by Engine. ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ ... │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/model.py:289 in validate │ │ │ │ 286 │ │ """%(validate_doc)s""" │ │ 287 │ │ return cast( │ │ 288 │ │ │ DataFrameBase[TDataFrameModel], │ │ ❱ 289 │ │ │ cls.to_schema().validate( │ │ 290 │ │ │ │ check_obj, head, tail, sample, random_state, lazy, inplace │ │ 291 │ │ │ ), │ │ 292 │ │ ) │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/model.py:242 in to_schema │ │ │ │ 239 │ │ ) │ │ 240 │ │ cls.__root_checks__ = df_custom_checks + df_registered_checks │ │ 241 │ │ │ │ ❱ 242 │ │ columns = cls._build_columns_index(cls.__fields__, cls.__checks__) │ │ 243 │ │ │ │ 244 │ │ kwargs = {} │ │ 245 │ │ if cls.__config__ is not None: │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/model.py:330 in │ │ _build_columns_index │ │ │ │ 327 │ │ │ │ │ │ f"'check_name' is not supported for {field_name}." │ │ 328 │ │ │ │ │ ) │ │ 329 │ │ │ │ │ │ ❱ 330 │ │ │ │ columns[field_name] = col_constructor( # type: ignore │ │ 331 │ │ │ │ │ dtype, │ │ 332 │ │ │ │ │ required=not annotation.optional, │ │ 333 │ │ │ │ │ checks=field_checks, │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/model_components.py:61 in │ │ to_column │ │ │ │ 58 │ │ name: str = None, │ │ 59 │ ) -> Column: │ │ 60 │ │ """Create a schema_components.Column from a field.""" │ │ ❱ 61 │ │ return self._to_schema_component( │ │ 62 │ │ │ dtype, │ │ 63 │ │ │ Column, │ │ 64 │ │ │ nullable=self.nullable, │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/model_components.py:51 in │ │ _to_schema_component │ │ │ │ 48 │ │ if self.dtype_kwargs: │ │ 49 │ │ │ dtype = dtype(**self.dtype_kwargs) # type: ignore │ │ 50 │ │ checks = self.checks + to_checklist(checks) │ │ ❱ 51 │ │ return component(dtype, checks=checks, **kwargs) # type: ignore │ │ 52 │ │ │ 53 │ def to_column( │ │ 54 │ │ self, │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/components.py:70 in __init__ │ │ │ │ 67 │ │ │ │ 68 │ │ See :ref:`here<column>` for more usage details. │ │ 69 │ │ """ │ │ ❱ 70 │ │ super().__init__( │ │ 71 │ │ │ dtype=dtype, │ │ 72 │ │ │ checks=checks, │ │ 73 │ │ │ nullable=nullable, │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/column_schema.py:52 in │ │ __init__ │ │ │ │ 49 │ │ :type nullable: bool │ │ 50 │ │ """ │ │ 51 │ │ │ │ ❱ 52 │ │ super().__init__( │ │ 53 │ │ │ dtype=dtype, │ │ 54 │ │ │ checks=checks, │ │ 55 │ │ │ coerce=coerce, │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/base/schema.py:39 in __init__ │ │ │ │ 36 │ │ drop_invalid_rows=False, │ │ 37 │ ): │ │ 38 │ │ """Abstract base schema initializer.""" │ │ ❱ 39 │ │ self.dtype = dtype │ │ 40 │ │ self.checks = checks │ │ 41 │ │ self.coerce = coerce │ │ 42 │ │ self.name = name │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/api/pyspark/column_schema.py:81 in dtype │ │ │ │ 78 │ def dtype(self, value: Optional[PySparkDtypeInputTypes]) -> None: │ │ 79 │ │ """Set the pyspark dtype""" │ │ 80 │ │ self._dtype = ( │ │ ❱ 81 │ │ │ pyspark_engine.Engine.dtype(value) if value else None │ │ 82 │ │ ) # pylint:disable=no-value-for-parameter │ │ 83 │ │ │ 84 │ def validate( │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/engines/pyspark_engine.py:125 in dtype │ │ │ │ 122 │ │ │ │ subst = "" │ │ 123 │ │ │ │ # You can manually specify the number of replacements by changing the 4t │ │ 124 │ │ │ │ data_type = re.sub(regex, subst, data_type, 0, re.MULTILINE) │ │ ❱ 125 │ │ │ return engine.Engine.dtype(cls, data_type) │ │ 126 │ │ except TypeError: # pylint:disable=try-except-raise # pragma: no cover │ │ 127 │ │ │ raise │ │ 128 │ │ │ │ /Users/fo2/c1s/worktree/pyspark_optional_fields/pandera/engines/engine.py:271 in dtype │ │ │ │ 268 │ │ try: │ │ 269 │ │ │ return registry.dispatch(data_type) │ │ 270 │ │ except (KeyError, ValueError): │ │ ❱ 271 │ │ │ raise TypeError( │ │ 272 │ │ │ │ f"Data type '{data_type}' not understood by {cls.__name__}." │ │ 273 │ │ │ ) from None │ │ 274 │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ TypeError: Data type '<class 'pyspark.sql.types.TimestampNTZType'>' not understood by Engine.Issue
The
pyspark_engine.pymodule was not prepared to processpyspark.sql.types.TimestampNTZType()as a Timestamp.Proposed solution
As this type is exclusive to Spark 3.4 and newer versions, an IF condition was needed to ensure that it's added only when pyspark >= 3.4 is being used, both in the main code as in tests:
Behavior when using PySpark 3.3, with common types only:
Behavior when using PySpark 3.4, with the new type:
Additional details
Version in use: 0.17.2
OS: MacOS 13.6 (M1)