DOC: Examples of monotonic/non-monotonic label-based indexing #12522
Conversation
nileracecrew
changed the title
| @@ -242,6 +242,57 @@ Label-based slicing conventions | |||
| Non-monotonic indexes require exact matches | |||
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |||
|
|
|||
| If the index of a ``DataFrame`` is montonically increasing or decreasing, then the bounds | |||
There was a problem hiding this comment.
typo in 'montonically'
|
@nileracecrew Thanks for this nice contribution to the docs! (that title was indeed missing an explanation ;-)) It is certainly good to keep it here, but we maybe should also consider to include a similar explanation in the indexing docs itself. |
|
You're welcoem @jorisvandenbossche. Seems like this behavior is most likely to show up when you have other problems (in my case corrupted data, as well as an in another issue I found), but it is surprising. I made the suggested edits and re-committed. |
|
instead a empty frame use a range for number (range(6)) |
|
Sorry, I don't quite understand your meaning. Do you mean fill the example DataFrame with numbers, or use range(6) as an index instead of the empty DataFrame example, or something else? |
|
fill with numbers |
|
@nileracecrew : If I'm not mistaken (unless the tests do check these files), since your fixes are purely documentation related, you should put [ci skip] in your commit message because you aren't changing any code. Every time you make a change to your PR, you are adding another build to Travis (you currently have five). Unless these |
| If the index of a ``DataFrame`` is monotonically increasing or decreasing, then the bounds | ||
| of a label-based slice can be outside the range of the index, much like slice indexing a | ||
| normal Python ``list``. | ||
|
|
There was a problem hiding this comment.
hmm actually what I mean is we need to more references to this section as this is certainly easy to miss. maybe can do that in antother PR though.
| >>> df_monotonic_index = pd.DataFrame(index=[2,3,3,4,5], columns=['data'], data=range(5)) | ||
|
|
||
| # no rows 0 or 1, but still returns rows 2, 3 (both of them), and 4: | ||
| >>> df_monotonic_index.loc[0:4, :] |
There was a problem hiding this comment.
actually make these all ipython blocks rather than code-blocks.
There was a problem hiding this comment.
In this case it makes sense to use code blocks I think, as below there are some error messages shown (and they don't need to be shown fully).
But, maybe use IPython prompts for uniformity with the rest of the document?
|
pls squash when you update |
|
Is there a quick way to render a single .rst file to see the output? GitHub doesn't render the ipython blocks. |
|
sorry, didn't mean to push that here |
|
thanks! |
Just ran into this gotcha and almost filed it as a bug report. Added a few examples to fill in the section header.