Jedi18 commented on code in PR #14781:
URL: https://github.com/apache/arrow/pull/14781#discussion_r1042278933
##########
python/pyarrow/tests/test_table.py:
##########
@@ -2204,3 +2204,100 @@ def test_table_cast_invalid():
table = pa.table({'a': [None, 1], 'b': [False, True]})
assert table.cast(new_schema).schema == new_schema
+
+
+def test_table_sort():
+ tab = pa.Table.from_arrays([
+ pa.array([5, 7, 7, 35], type=pa.int64()),
+ pa.array(["foo", "car", "bar", "foobar"])
+ ], names=["a", "b"])
+
+ sorted_tab = tab.sort_by([("a", "descending")])
+ sorted_tab_dict = sorted_tab.to_pydict()
+ assert sorted_tab_dict["a"] == [35, 7, 7, 5]
+ assert sorted_tab_dict["b"] == ["foobar", "car", "bar", "foo"]
+
+ sorted_tab = tab.sort_by([("a", "ascending")])
+ sorted_tab_dict = sorted_tab.to_pydict()
+ assert sorted_tab_dict["a"] == [5, 7, 7, 35]
+ assert sorted_tab_dict["b"] == ["foo", "car", "bar", "foobar"]
+
+
+def test_record_batch_sort():
+ rb = pa.RecordBatch.from_arrays([
+ pa.array([7, 35, 7, 5], type=pa.int64()),
+ pa.array([4, 1, 3, 2], type=pa.int64()),
+ pa.array(["foo", "car", "bar", "foobar"])
+ ], names=["a", "b", "c"])
+
+ sorted_rb = rb.sort_by([("a", "descending"), ("b", "descending")])
+ sorted_rb_dict = sorted_rb.to_pydict()
+ assert sorted_rb_dict["a"] == [35, 7, 7, 5]
+ assert sorted_rb_dict["b"] == [1, 4, 3, 2]
+ assert sorted_rb_dict["c"] == ["car", "foo", "bar", "foobar"]
+
+ sorted_rb = rb.sort_by([("a", "ascending"), ("b", "ascending")])
+ sorted_rb_dict = sorted_rb.to_pydict()
+ assert sorted_rb_dict["a"] == [5, 7, 7, 35]
+ assert sorted_rb_dict["b"] == [2, 3, 4, 1]
+ assert sorted_rb_dict["c"] == ["foobar", "bar", "foo", "car"]
+
+ # test multi-key record batch sorter (> 8 sort keys)
Review Comment:
I do see c++ unit tests with more than 8 keys so yes I guess it's safe to
assume both implementations are tested. I agree that we could probably remove
these extra tests since the python tests should not be concerned with the
details of the internal cpp implementation.
--
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]
