Re-implemented parametrization of test_frame_from_json_to_json #28510
Conversation
WillAyd
added
IO JSON
WillAyd
changed the title
|
I'll take a look. Thanks for separating this out. |
| ) | ||
|
|
||
| expected = df.copy() | ||
| expected = expected.assign(**expected.select_dtypes("number").astype(np.int64)) |
There was a problem hiding this comment.
I dont use assign or select_dtypes very often. Is this just casting to int64 for columns A and B?
There was a problem hiding this comment.
Yes that's correct. Happy to change to another way of casting if you prefer
There was a problem hiding this comment.
I think df[["A", "B"]] = df[["A", "B"]].astype(np.int64) would be clearer. Might just be me.
| if orient == "records" or orient == "values": | ||
| expected = expected.reset_index(drop=True) | ||
| if orient == "values": | ||
| expected.columns = range(len(expected.columns)) |
There was a problem hiding this comment.
re-iterating request to share code for ~344-360. As is I have to look at each version for "is this subtly different and if so why"
There was a problem hiding this comment.
Hmm I don't know I fully understand what you are asking for. This particular test lines up pretty well to the left (lines 463 - 472) but is more explicit about the expected shape of the result. This idiom of resetting particular axes for values and records is used throughout the module (probably worth a dedicated function, but leaving to a follow up)
There was a problem hiding this comment.
This idiom of resetting particular axes for values and records is used throughout the module
Yah, thats what I'm asking for. OK for follow-up.
| self.empty_frame, check_index_type=False, check_column_type=False | ||
| @pytest.mark.parametrize("convert_axes", [True, False]) | ||
| @pytest.mark.parametrize("numpy", [True, False]) | ||
| def test_roundtrip_timestamp(self, orient, convert_axes, numpy): |
There was a problem hiding this comment.
LGTM, corresponds to 452-453
|
|
||
| @pytest.mark.parametrize("convert_axes", [True, False]) | ||
| @pytest.mark.parametrize("numpy", [True, False]) | ||
| def test_roundtrip_empty(self, orient, convert_axes, numpy): |
There was a problem hiding this comment.
LGTM, corresponds to 447-449
| @pytest.mark.parametrize("dtype", [False, np.int64]) | ||
| @pytest.mark.parametrize("convert_axes", [True, False]) | ||
| @pytest.mark.parametrize("numpy", [True, False]) | ||
| def test_roundtrip_intframe(self, orient, convert_axes, numpy, dtype): |
There was a problem hiding this comment.
func LGTM. corresponds to 417-418
| @pytest.mark.parametrize("dtype", [False, float]) | ||
| @pytest.mark.parametrize("convert_axes", [True, False]) | ||
| @pytest.mark.parametrize("numpy", [True, False]) | ||
| def test_roundtrip_simple(self, orient, convert_axes, numpy, dtype): |
There was a problem hiding this comment.
func LGTM. corresponds to 413-415
|
Thanks @WillAyd |
This piece was broken off of #27838 as it made the diff much larger, so hopefully easier to digest on its own.
As is, parametrization here has brought up a lot of rough edges which are responsible for some of the complexity. These are noted with TODOs and summarized as follows (save Py35 issues, which aren't worth addressing at this point):
numpy=False(default) andorient="index"np.int64roundtrips asnp.int32(maybe not an issue?)orient="split"does not preserve strings in the index if those strings are numeric, though it should be able toconvert_axesmay have surprising behavior when dealing with empty DataFramesNot all of these are the same priority, but figure worth leaving as follow ups