Add informal definition of "label" in note #27005
Conversation
The term "label" is used extenisively, but (AFAICT) it is never explained what a "label" is. Is an integer offset a label? Do labels apply to rows and columns, multi-indexes? What's the relation between an "index" "axis" and "label" ... etc.
Codecov Report
@@ Coverage Diff @@
## master #27005 +/- ##
==========================================
- Coverage 91.99% 91.99% -0.01%
==========================================
Files 180 180
Lines 50774 50774
==========================================
- Hits 46712 46708 -4
- Misses 4062 4066 +4
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #27005 +/- ##
==========================================
- Coverage 91.99% 91.99% -0.01%
==========================================
Files 180 180
Lines 50774 50774
==========================================
- Hits 46712 46708 -4
- Misses 4062 4066 +4
Continue to review full report at Codecov.
|
|
I find this note a little confusing. Would just mentioning that a |
|
I agree. It is too verbose, and there is already too many notes and warnings (which I tend to have a habit of not reading anyway :)). I will re-word. |
There was a problem hiding this comment.
Hmm still not sure this is it but cc @datapythonista in case he has any insight to offer
| @@ -6,17 +6,18 @@ | |||
| Indexing and Selecting Data | |||
| *************************** | |||
|
|
|||
| The axis labeling information in pandas objects serves many purposes: | |||
| In pandas, data structure axes such as the rows and columns of a DataFrame may have | |||
There was a problem hiding this comment.
To be clear, an Index (which is used for selection along an axis) does have labels. This is a main thing that separates pandas from libraries like NumPy
There was a problem hiding this comment.
Yes I know. It's actually quite challenging to unpack axis, index, labels, positional indexation via integers in a sentence or two. An index really is labels. I found it an improvement over "axis labeling information ... serves many purposes", by being less nebulous and vague.
|
I agree that we should be more explicit on the terms we used. I think we discussed in the past about having a glossary, which I think it's the right way to address this. I could only find this related issue: #21784 I don't think this PR is helping much. I'll close this, and open an issue for the glossary, feel free to work on that. |
|
Sounds good. May be loosely intertwined with what we are defining in pandas._typing @sam-at-github thanks for the PR. To the point above though I think we'd rather make a more comprehensive change than what is offered here. Would certainly value input if you'd like to put thoughts into the other issue |
|
Sorry, re-reading my last comment, just realized that probably didn't sound really nice... Not my intention. @sam-at-github thanks for your contribution, and it really helps, since this is very helpful feedback. I opened #27019 since I think a glossary will make it easier to address what you're addressing here. Feel free to continue the work you started here, but I think it makes more sense to close this, and continue in a PR that addresses that issue. Thanks! |
|
Sounds like perfect is the enemy of the good around here. Good luck with glossary. |
The term "label" is used extenisively, but (AFAICT) it is never explained what a "label" is. Is an integer offset a label? Do labels apply to rows and columns, multi-indexes? What's the relation between an "index" "axis" and "label" ... etc.
git diff upstream/master -u -- "*.py" | flake8 --diff