BUG: Series constructor with category dtype does not raise with unknown categories

[WillAyd]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

Assignment fails as expected:


>>> cat = pd.CategoricalDtype(["foo", "bar", "baz"])
>>> ser = pd.Series(["foo", "baz", "baz"], dtype=cat)
>>> ser.iloc[2] = "qux"
TypeError: Cannot setitem on a Categorical with a new category (qux), set the categories first


But the constructor just maps this to a missing value:
```python
>>> pd.Series(["foo", "baz", "baz", "qux"], dtype=cat)
Out[13]: 
0    foo
1    baz
2    baz
3    NaN
dtype: category
Categories (3, object): ['foo', 'bar', 'baz']

### Issue Description

The Series constructor allows you to use values that are not part of the categorical values, mapping them to missing

### Expected Behavior

I think the constructor should fail just like assignment does

### Installed Versions

main

[WillAyd]

This might be another one of those constructors quirks @Aloqeely and I were chatting about on Slack recently

[WillAyd]

Though our documentation does rely on this behavior https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.cat.set_categories.html

Maybe it is considered a feature?

[jbrockmendel]

Xref #40996

[WillAyd]

Ah thanks - this is a dupe then

[jbrockmendel]

I like your framing of it as a bug, would be happy to keep this one and close the other.

…

On Mon, Jul 8, 2024 at 6:52 AM William Ayd ***@***.***> wrote: Ah thanks - this is a dupe then — Reply to this email directly, view it on GitHub <#59204 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5UM6A5YGXYSAQSBFNWHQLZLKKS7AVCNFSM6AAAAABKPYTNAGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMJUGEZTQNBUGI> . You are receiving this because you commented.Message ID: ***@***.***>