jorisvandenbossche commented on a change in pull request #11008:
URL: https://github.com/apache/arrow/pull/11008#discussion_r712777558
##########
File path: python/pyarrow/tests/test_dataset.py
##########
@@ -3231,6 +3224,85 @@ def test_write_dataset_partitioned(tempdir):
partitioning=partitioning_schema)
+def test_dictionary_partitioning_outer_nulls_raises(tempdir):
+ table = pa.table({'a': ['x', 'y', None], 'b': ['x', 'y', 'z']})
+ part = ds.partitioning(
+ pa.schema([pa.field('a', pa.string()), pa.field('b', pa.string())]))
+ with pytest.raises(pa.ArrowInvalid):
+ ds.write_dataset(table, tempdir, format='parquet', partitioning=part)
+
+
+def test_write_dataset_with_field_names(tempdir):
+ table = pa.table({'a': ['x', 'y', None], 'b': ['x', 'y', 'z']})
+
+ ds.write_dataset(table, tempdir, format='parquet',
+ partitioning=["b"])
+
+ load_back = ds.dataset(tempdir, partitioning=["b"])
+ files = load_back.files
+ partitioning_dirs = {
+ str(pathlib.Path(f).relative_to(tempdir).parent) for f in files
+ }
+ assert partitioning_dirs == {"x", "y", "z"}
+
+ load_back_table = load_back.to_table()
+ assert load_back_table.equals(table)
+
+
+def test_write_dataset_with_field_names_hive(tempdir):
+ table = pa.table({'a': ['x', 'y', None], 'b': ['x', 'y', 'z']})
+
+ ds.write_dataset(table, tempdir, format='parquet',
+ partitioning=["b"], partitioning_flavor="hive")
+
+ load_back = ds.dataset(tempdir, partitioning="hive")
+ files = load_back.files
+ partitioning_dirs = {
+ str(pathlib.Path(f).relative_to(tempdir).parent) for f in files
+ }
+ assert partitioning_dirs == {"b=x", "b=y", "b=z"}
+
+ load_back_table = load_back.to_table()
+ assert load_back_table.equals(table)
+
+
+def test_write_dataset_with_scanner(tempdir):
+ table = pa.table({'a': ['x', 'y', None], 'b': ['x', 'y', 'z'],
+ 'c': [1, 2, 3]})
+
+ ds.write_dataset(table, tempdir, format='parquet',
+ partitioning=["b"])
+
+ dataset = ds.dataset(tempdir, partitioning=["b"])
+
+ with tempfile.TemporaryDirectory() as tempdir2:
+ ds.write_dataset(dataset.scanner(columns=["b", "c"]), tempdir2,
+ format='parquet', partitioning=["b"])
+
+ load_back = ds.dataset(tempdir2, partitioning=["b"])
+ load_back_table = load_back.to_table()
+ assert load_back_table.to_pydict() == {'b': ['x', 'y', 'z'],
+ 'c': [1, 2, 3]}
Review comment:
(the tests you added above are using the `equals(..)` pattern ..)
You're right that the equals method doesn't give much useful information if
not equal, but I think we should then solve that by adding a test utility like
`assert_table_equals` or so, instead of having to repeat the data as a dict.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
