add conditional to check ndim
#63058
Conversation
|
Please review this. |
|
Does this PR address an open issue? |
Not any issue which I know of. But, I thought that it should raise an error. So, I made the PR. |
|
Sorry for being pedantic about this, but since this is a change in behavior, I think it's best to:
|
I will do what you said. But, I am still learning. I do not know to create tests. Please help me with that. |
|
Issue created for this PR #63112 |
|
I think that |
I will do that. Should I make a new PR for that? |
It should be done in this one. |
Thank you for helping. I will create the tests. |
|
pre-commit failed because the max characters allowed in a line exceeded. I did not know about the max allowed characters. Sorry for that. |
pandas/tests/arrays/test_array.py
Outdated
|
|
||
| @pytest.mark.parametrize("data", [np.array(0)]) | ||
| def test_nd_raises(data): | ||
| @pytest.mark.parametrize("data,dtype", [(np.array(0),"int64"),(np.array([[1, 2], ["a", "b"]]),object)]) |
There was a problem hiding this comment.
You can fix the formatting with pre-commit. The docs have a section about it. It's also good to create a development environment.
To fix the formatting, setup pre-commit and run
pre-commit run --files pandas/tests/arrays/test_array.py
pandas/core/construction.py
Outdated
| if data.ndim != 1: | ||
| raise ValueError("NumpyExtensionArray must be 1-dimensional") |
There was a problem hiding this comment.
The test shows that the verification shouldn't be put here. I think the correct place is pandas/core/arrays/numpy_.py
Change dimensionality check for NumpyExtensionArray.
|
Tried making changes to |
2 participants
Description
It checks the value of
ndimwhen the argument fordatais annp.ndarrayinpandas.array(data, dtype=None, copy=True)Reason
Raises an error because the array is 2-D
Does not raise an error in this case(maybe, because the value is now a string). But, it should raise an error as only 1-D arrays can be passed as arguments.
resolves #63112