PEP 634 specifies that

"A mapping pattern may not contain duplicate key values. (If all key patterns are literal patterns this is considered a syntax error; otherwise this is a runtime error and will raise ValueError.)"

but this is not what happens with the latest release:

Python 3.10.0b3 (tags/v3.10.0b3:865714a, Jun 17 2021, 20:39:25) [MSC v.1929 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> x = {'a': 1}
>>> match x:
...  case {'a': 1, 'a': 2}: # (A)
...   print('ok')
...
>>> x = {'a': 3}
>>> match x:
...  case {'a': 1, 'a': 2}: # (B)
...   print('ok')
...
>>> x = {'a': 1, 'b': 2}
>>> match x:
...  case {'a': 1, 'a': 2}: # (C)
...   print('ok')
...
Traceback (most recent call last):
  File "<stdin>", line 2, in <module>
ValueError: mapping pattern checks duplicate key ('a')
>>>

If I understand the PEP correctly, all these examples should raise a SyntaxError for the line

case {'a': 1, 'a': 2}:

since all key patterns are literal patterns, and the key 'a' is duplicated.

Cases (A) where the subject matches one of the key-value patterns, and (B) when it doesn't, fail without raising SyntaxError.

Case (C) where one of the keys in the subject is not present in the mapping pattern raises a ValueError at runtime instead of SyntaxError.

This behaviour is tested in test_patma.py:

    def test_patma_251(self):
        x = {"a": 0, "b": 1}
        w = y = z = None
        with self.assertRaises(ValueError):
            match x:
                case {"a": y, "a": z}:
                    w = 0
        self.assertIs(w, None)
        self.assertIs(y, None)
        self.assertIs(z, None)

but this doesn't seem compliant with the specification.

BTW, it's not clear to me why the SyntaxError should be limited to the case when all keys are literal patterns; it could be raised whenever a literal pattern is repeated, even when there are value patterns or a double-star pattern, like in

case {'a': 1, 'a': 2, c.imag, **rest}:

Nice find! In my opinion, we should do two things here:

- Update PEP 638 to specify that a SyntaxError is raised "if *any duplicate* key patterns are literal patterns". This was absolutely our original intention... I think the nuances were just lost in the final phrasing. I can take care of that PR.

- Update the compiler to raise SyntaxErrors for duplicate literal keys. It should be as simple as updating the first loop in compiler_pattern_mapping to build a set containing the values of any literal keys and raise on duplicates (and adding/updating tests, of course). We'll want to make sure we have test coverage of edge cases like {0: _, False: _}, {0: _, 0.0: _}, {0: _, -0: _}, {0: _, 0j: _}, etc.

Since you found this, you get first dibs on a PR. Otherwise, I have a few first-time contributors who would probably be interested.

Sorry, that should be PEP 634, not 638.

Agreed with everything that Brandt said. These wording issues are subtle!
Note in particular that the original wording implies that {‘x’: 1, ‘x’: 1,
z: 1} would be a runtime error, but it clearly should be a compile time
error.--
--Guido (mobile)

Sorry, I don't know C so I can't write a PR for this change.

@brandtbucher, I think that my PR implements the solution you've described here. What do you think?

Thanks for the patch @jack__d! I'll try to find time to review it today.

I do wish you would have coordinated with me here before writing it, though. I'd already begun working on a patch with a few new contributors yesterday, as I mentioned in my comment.

In the future, please check in with those involved in the PR discussion first (*especially* if the issue is already "assigned"). We're doing this for fun, too. :)

@brandtbucher, I'm sorry for the miscommunication. I started working on this patch at the beginning of the week after message 397215... I'm a new contributor too, and I wasn't sure if I would be able to make something that worked, so I just kept my mouth shut. Then, all of a sudden, it came together and I had a (mostly) working patch.

I'm going to incorporate your review now; thank you for the feedback, and I'm very sorry for any toes I may have stepped on.

New changeset 2693132292b2acf381ac6fa729bf3acf41d9d72b by Jack DeVries in branch 'main':
bpo-44589: raise a SyntaxError when mapping patterns have duplicate literal keys (GH-27131)
https://github.com/python/cpython/commit/2693132292b2acf381ac6fa729bf3acf41d9d72b

New changeset 016af14d93cfba43e7a95721a97fa954c534af8e by Miss Islington (bot) in branch '3.10':
[3.10] bpo-44589: raise a SyntaxError when mapping patterns have duplicate literal keys (GH-27131) (GH-27157)
https://github.com/python/cpython/commit/016af14d93cfba43e7a95721a97fa954c534af8e

Is something left here?

The PR and backport to 3.10 have both been merged, so I think this issue can be closed.