Skip to content

[SPARK-30091][SQL][Python] Document mergeSchema option directly in the PySpark Parquet APIs #26730

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
nchammas wants to merge 2 commits into from
Closed

[SPARK-30091][SQL][Python] Document mergeSchema option directly in the PySpark Parquet APIs #26730

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
da50864
document mergeSchema directly in API
nchammas Dec 2, 2019
33770b5
Merge branch 'master' of https://github.com/apache/spark into parquet…
nchammas Dec 4, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 8 additions & 6 deletions python/pyspark/sql/readwriter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -300,18 +300,20 @@ def table(self, tableName):
return self._df(self._jreader.table(tableName))

@since(1.4)
def parquet(self, *paths):
Copy link
Contributor Author

@nchammas nchammas Dec 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Side question for you @HyukjinKwon: The *paths parameter bothers me a bit. None of the other load methods use this pattern, and the streaming version of parquet() doesn't use it either. How would you feel about a separate PR changing this to paths? I suppose the 3.0 release would be our chance to do it, since it changes the API.

Copy link
Member

@HyukjinKwon HyukjinKwon Dec 2, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this *paths allows a consistent way with Scala and Java side (def parquet(paths: String*)). So I think technically it's more correct to support *paths.

In case of streaming, it's also matched to Scala / Java side def parquet(path: String).

Maybe we should introduce keyword-only argument (as you said earlier somewhere) after completely dropping Python 2 in Spark 3.1. ... I am not sure about this yet.

"""Loads Parquet files, returning the result as a :class:`DataFrame`.
def parquet(self, *paths, **options):
"""
Loads Parquet files, returning the result as a :class:`DataFrame`.

You can set the following Parquet-specific option(s) for reading Parquet files:
* ``mergeSchema``: sets whether we should merge schemas collected from all \
Parquet part-files. This will override ``spark.sql.parquet.mergeSchema``. \
The default value is specified in ``spark.sql.parquet.mergeSchema``.
:param mergeSchema: sets whether we should merge schemas collected from all
Parquet part-files. This will override ``spark.sql.parquet.mergeSchema``.
The default value is specified in ``spark.sql.parquet.mergeSchema``.

>>> df = spark.read.parquet('python/test_support/sql/parquet_partitioned')
>>> df.dtypes
[('name', 'string'), ('year', 'int'), ('month', 'int'), ('day', 'int')]
"""
mergeSchema = options.get('mergeSchema', None)
self._set_opts(mergeSchema=mergeSchema)
return self._df(self._jreader.parquet(_to_seq(self._spark._sc, paths)))

@ignore_unicode_prefix
Expand Down
15 changes: 8 additions & 7 deletions python/pyspark/sql/streaming.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -526,22 +526,23 @@ def orc(self, path):
raise TypeError("path can be only a single string")

@since(2.0)
def parquet(self, path):
"""Loads a Parquet file stream, returning the result as a :class:`DataFrame`.

You can set the following Parquet-specific option(s) for reading Parquet files:
* ``mergeSchema``: sets whether we should merge schemas collected from all \
Parquet part-files. This will override ``spark.sql.parquet.mergeSchema``. \
The default value is specified in ``spark.sql.parquet.mergeSchema``.
def parquet(self, path, mergeSchema=None):
"""
Loads a Parquet file stream, returning the result as a :class:`DataFrame`.

.. note:: Evolving.

:param mergeSchema: sets whether we should merge schemas collected from all
Parquet part-files. This will override ``spark.sql.parquet.mergeSchema``.
The default value is specified in ``spark.sql.parquet.mergeSchema``.

>>> parquet_sdf = spark.readStream.schema(sdf_schema).parquet(tempfile.mkdtemp())
>>> parquet_sdf.isStreaming
True
>>> parquet_sdf.schema == sdf_schema
True
"""
self._set_opts(mergeSchema=mergeSchema)
if isinstance(path, basestring):
return self._df(self._jreader.parquet(path))
else:
Expand Down