DOC: add explanation of dtype to parameter values for np.append - #26303

@rkern @seberg I took a first stab at it. I need your guidance here.

install pkg-config is failing. I dont think its happening due to my code changes.

all right. tests have passed now.

This PR is ready for review

cc @rkern @seberg

Please add [skip azp][skip cirrus][skip actions] to the commit message to only run the circleci documentation CI job.

Please add [skip azp][skip cirrus][skip actions] to the commit message to only run the circleci documentation CI job.

Thank you for pointing. Done.

@mattip Please let me know if any more changes need to be done.

I think this does not actually address the issue that prompted the PR, which was that calling np.append(np.array([1]), np.array([])) will result in a float64 array, not an int one. The change here makes sense, but could be added to pretty much every function that mixes two ndarrays. The special thing in the issue is that an empty array will have dtype=float64. Can you think of a way to document that special case? Maybe adding a Note, or an Example?

Thanks @tuhinsharma121