BUG: compute.use_numexpr option not respected #42668
Conversation
saehuihwang
commented
Jul 22, 2021
saehuihwang
commented
- closes Dataframe.query uses numexpr even when 'compute.use_numexpr' is False #32556
- tests added / passed
- Ensure all linting tests pass, see here for how to run them
- whatsnew entry
| @@ -1865,6 +1865,16 @@ def test_invalid_engine(): | |||
| pd.eval("x + y", local_dict={"x": 1, "y": 2}, engine="asdf") | |||
|
|
|||
|
|
|||
| @td.skip_if_no_ne | |||
There was a problem hiding this comment.
can you parameterize and text for both True/False here.
| # GH 32556 | ||
| from pandas.core.computation.eval import _check_engine | ||
|
|
||
| with pd.option_context("compute.use_numexpr", False): |
There was a problem hiding this comment.
can you add a test similar to #32556 (comment)
that we can check from a user perspective that this is working.
jreback
added
Algos
|
also look thru the referenced issue for any other things that we can test with this. it is quite trick to make sure this is working propertly. |
| if use_numexpr: | ||
| assert result == "numexpr" | ||
| else: | ||
| assert result == "python" |
There was a problem hiding this comment.
logic (e.g. if/else statements) in tests aren't ideal (https://testing.googleblog.com/2014/07/testing-on-toilet-dont-put-logic-in.html), how about something like
@pytest.mark.parametrize(
("use_numexpr", "expected"),
(
(True, "numexpr"),
(False, "python"),
)
)
def test_numexpr_option_respected(use_numexpr, expected):
and then, in the body, just assert result == expected?
There was a problem hiding this comment.
This is good to know! Thank you for linking the resource. Added!
|
lgtm. over to you @MarcoGorelli |
|
thanks @saehuihwang |
