BUG: fixes weekday for dates before 1752

[mcgeestocks]

• closes BUG: weekday incorrect for dates before 0000-03-01 #53738 (Replace xxxx with the GitHub issue number)
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

[mcgeestocks]

Approach was to switch to implementing Gauss's algorithm for weekday. Ended up being quite interesting.

[MarcoGorelli]

nice work!

just got a couple of requests:

• on the unit test side, could you add a couple of negative dates please?
• could we also add a parametric test using hypothesis please, in which we compare the output against the standard library output? feel free to reach out if you're not sure how to do this

[mcgeestocks]

Thanks @MarcoGorelli!

Haven't used Hypothesis before by tried to follow the patterns elsewhere in the codebase.
Also by "standard library" I assumed you were referring to datetime. Let me know if I missed the mark any what you were looking for.

[MarcoGorelli]

yup, that's right! one way to do it could be to make a dict with

{0: 'Mon', 1: 'Tue', ...}

and then generate a random datetime datetime, and compare Timestamp(datetime).weekday with datetime.strftime('%a')

something like that

[mcgeestocks]

Sorry @MarcoGorelli did you see this push ca826d8 ?

I had already added the Hypothesis driven test before my pervious note, sorry for the confusion.

[MarcoGorelli]

nice one! just got a comment on the parametric test

[MarcoGorelli]

generally looks good, just got quite a few minor comments

[MarcoGorelli]

Looks good to me, thanks @mcgeestocks !

I've just moved this to the 2.0.3 milestone, as technically it's a regression introduced by version 2.0.0 (before, .dt.weekday was always correct, as timestamps were only supported between about 1700 and 2300)

EDIT: requesting changes for now, see below

[MarcoGorelli]

CI failure is unrelated, but probably safer to wait til that's addressed before merging

[MarcoGorelli]

Although this looks correct, the implementation looks a bit different to what's linked to in https://en.wikipedia.org/wiki/Determination_of_the_day_of_the_week#Gauss's_algorithm

could you explain what the various letters you've introduced (e.g. e) refer to please?

EDIT: seems like you've used the implementation from http://berndt-schwerdtfeger.de/wp-content/uploads/pdf/cal.pdf, which looks good. All good then, I'll add a note linking to that

[MarcoGorelli]

Right, I've noted the source of the implementation, have simplified slightly, and have expanded the parametric test to include negative dates (by comparing against numpy)

Very confident about this now - will merge today unless there's any objections, would be good to get this fix into 2.0.3

[lumberbot-app]

Owee, I'm MrMeeseeks, Look at me.

There seem to be a conflict, please backport manually. Here are approximate instructions:

1. Checkout backport branch and update it.

git checkout 2.0.x
git pull

2. Cherry pick the first parent branch of the this PR on top of the older branch:

git cherry-pick -x -m1 20530b1f45045bfb95b0a79cda70fa7b804c9a6a

3. You will likely have some merge/cherry-pick conflict here, fix them and commit:

git commit -am 'Backport PR #53795: BUG: fixes weekday for dates before 1752'

4. Push to a named branch:

git push YOURFORK 2.0.x:auto-backport-of-pr-53795-on-2.0.x

5. Create a PR against branch 2.0.x, I would have named this PR:

"Backport PR #53795 on branch 2.0.x (BUG: fixes weekday for dates before 1752)"

And apply the correct labels and milestones.

Congratulations — you did some good work! Hopefully your backport PR will be tested by the continuous integration and merged soon!

Remember to remove the Still Needs Manual Backport label once the PR gets merged.

If these instructions are inaccurate, feel free to suggest an improvement.