QueryDict __getitem__ returns an empty list when the value is an empty list

[Nguyễn Hồng Quân]

Update: forum post: ​https://forum.djangoproject.com/t/change-querydict-getitem-in-case-of-empty-list/36522

Though ​documentation says that

QueryDict.__getitem__(key)
Returns the value for the given key. If the key has more than one value, it returns the last value.

It returns a list when the raw value is an empty list:

>>> from django.http import QueryDict
>>> q = QueryDict('a=1', mutable=True)
>>> q['a']
'1'

>>> q.setlist('b', [])
>>> q['b']
[]

which surprises user and it even not mentioned in documentation.

Could we change this behavior? I know that we don't have a perfect solution, but returning None in this case is less bad than empty list [], because it is easier to annotate type.

• If returns None, we can annotate type as str | None.
• If returns [], Python doesn't have an exact type for empty list, and django-stubs has to annotate as str | list[object] which is quite broader than the actual value (empty list).

[Sarah Boyce]

Could we change this behavior? I know that we don't have a perfect solution, but returning None in this case is less bad than empty list [], because it is easier to annotate type.

It's been this way a very long time, I'm not sure what the implications of changing it would be (in terms of compatibility). I also think changing it for easier type annotations is not a strong argument here. You could raise this topic on the ​Django forum and get others opinions if you wish to change it.

That being said, I think a docs adjustment is required (note that the docstring states Return the last data value for this key, or [] if it's an empty list; raise KeyError if not found.).
Accepting on that basis.

[jburns6789]

Ok, looking over your suggestion,
I'm trying to figure out what I need to do, here is the doc section for QuerySet

​https://docs.djangoproject.com/en/5.1/ref/request-response/#django.http.QueryDict

QueryDict.getitem(key)¶
Returns the value for the given key. If the key has more than one value, it returns the last value. Raises django.utils.datastructures.MultiValueDictKeyError if the key does not exist. (This is a subclass of Python’s standard KeyError, so you can stick to catching KeyError.)

I need to state when a QueryDictkey returns [ ], the said key is associated with a empty list?

[Sarah Boyce]

Replying to jburns6789:

QueryDict.getitem(key)¶
Returns the value for the given key. If the key has more than one value, it returns the last value. Raises django.utils.datastructures.MultiValueDictKeyError if the key does not exist. (This is a subclass of Python’s standard KeyError, so you can stick to catching KeyError.)

I need to state when a QueryDictkey returns [ ], the said key is associated with a empty list?

Yes
I would be tempted to update this to be very similar to the doc string so maybe

docs/ref/request-response.txt

@@ -554 +554 @@
 
 .. method:: QueryDict.__getitem__(key)
 
-    Returns the value for the given key. If the key has more than one value,
-    it returns the last value. Raises
-    ``django.utils.datastructures.MultiValueDictKeyError`` if the key does not
-    exist. (This is a subclass of Python's standard :exc:`KeyError`, so you can
-    stick to catching ``KeyError``.)
+    Returns the last data value for the given key, or ``[]`` if it's an empty
+    list. Raises ``django.utils.datastructures.MultiValueDictKeyError`` if the
+    key does not exist. (This is a subclass of Python's standard
+    :exc:`KeyError`, so you can stick to catching ``KeyError``.)
 
 .. method:: QueryDict.__setitem__(key, value)

[jburns6789]

Hi, here is my link to the PR, someone else was reviewing, is there anything I need to in the ticket tracker?

[jburns6789]

​https://github.com/django/django/pull/18834

[jburns6789]

Updated requested information ​https://github.com/django/django/pull/18841