DOC: add Comparison with Excel

[afeld]

• closes Equivalent to the Excel fill handle #22993
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

Background

I teach a class on pandas for public policy students, and for many of them, spreadsheets are the only point of reference they have for working with tabular data. It would be very helpful to have (official) document comparing the two to point them to.

This is my first contribution to pandas and first time using reStructuredText, so feedback welcome. Thanks in advance!

TODOs

Making a running checklist to show what I've done already, and what else I plan to do. Hoping for some preliminary feedback (like is there still interest in having this page) before spending too much more time on it.

Happy to continue in this pull request until complete with all of them, or get this merged sooner than later and take care of the others in follow-up pull requests. Slight preference for the latter (some documentation being better than none, less to review at once, etc.), but open to whatever.

• start with what @rotuna did in DOC: Add Comparison with Excel documentation #23042
• add links to canonical Excel documentation, as previously requested
• fix for CI
• link from Getting Started pages
• incorporate structure from SAS/STATA comparison pages
  • Data structures
  • Data input / output
    • Mention read_excel()
  • Data operations
  • String processing
  • Merging
  • Missing data
  • GroupBy
  • Other considerations
• convert examples to use tips dataset
• add info about charting

Questions

• Which whatsnew file should I add to?
• I noticed that doc/source/_static is in the .gitignore, but there are files checked into that folder. Is that intentional? - CLN: remove duplicate banklist.html file #38739
• Some of the comparison documentation refers to "columns", while other refer to "Series". Is there a preference, or can they be used interchangeably?
• Since spreadsheet software is largely interchangeable/compatible, would it make sense to make the page more general as "Comparison to spreadsheets"?
• Thoughts about including slightly more subjective content, such as why one might want to use spreadsheets vs. pandas?

[MarcoGorelli]

Thanks @afeld - I built this locally and it generally looks good

To answer some questions:

Which whatsnew file should I add to?

It shouldn't be necessary to add one, that's usually for new features / bug and regression fixes

Since spreadsheet software is largely interchangeable/compatible, would it make sense to make the page more general as "Comparison to spreadsheets"?

I think so, perhaps it could be "Comparison with spreadsheets (e.g. Excel)"?

EDIT

I don't think this needs a whatsnew note

[MarcoGorelli]

nice!

[jreback]

@afeld can you post a rendered picture here of the new docs page here

[jreback]

@afeld looks great. ideally if you can add the excel (and pandas) references. am happy to merge (and can keep adding things in later PRs).

[jreback]

should also indicate that the row labels themselves are akin to the default RangeIndex

[afeld]

Added a mention of RangeIndex below. That work?

[jreback]

it might be useful to include .png's here if it helps explain material

[afeld]

Sure, will put something together. Ok as a follow-up?

[jreback]

sure

[afeld]

Screenshot of the new page

Let me know what you think! Nudge on the questions up top. I know there are a lot of commits in here; let me know if you want me to squash.

More I want to do with the page, but hoping this is close to being merge-able.

[afeld]

I wasn't quite sure whether to use the ~ in the class/method/function references, so let me know if you want that changed.

[jreback]

this is fine

[afeld]

I copied this code from another file and kept it consistent, but it wasn't clear why the URL was being passed in as a tuple rather than a single string.

[MarcoGorelli]

this isn't a tuple, it's still a single string

>>> url = (
...     'a'
...     'b'
... )
>>> url
'ab'

I presume it's done like this to keep it from getting too long

[afeld]

Ah, I missed that there wasn't a comma! Wasn't aware of implicit string concatenation. Thanks!

[jreback]

this is fine

[jreback]

thanks @afeld very nice. built docs will appear here: https://pandas.pydata.org/pandas-docs/dev/ (may take a little for the CI to do it). happy to have a followup as discussed.

[jreback]

minor point: https://pandas.pydata.org/pandas-docs/dev/getting_started/comparison/comparison_with_spreadsheets.html#pivot-tables

you can add margins=True to have the exact excel pivot table behavior you are showing in the png.