Created attachment 174650 [details]
Demo file with filtering error

The sample file contains the data needed to reproduce the error.
The custom utility correctly performs the case-insensitive duplicate removal task.

here is an ambiguity in the feature description that doesn't allow to clearly see if this is a bug or not. The question assumes that "Case sensitive" option of the filters applies to the "No duplications"; but help [1] does not state that; it only tells that (1) Case sensitive "Distinguishes between uppercase and lowercase letters when filtering the data", and (2) No duplication "Excludes duplicate rows in the list of filtered data". So the (1) is only said to apply to the filtering conditions - not to duplication finding? On one hand, this may be seen as "works as designed"; on the other hand, it's likely that users would expect the different behavior.

It's interesting what other spreadsheet applications do in this case.

Adding Eike to CC; also asking advice of UX team.

[1] https://help.libreoffice.org/7.2/en-US/text/scalc/01/12040201.html?&DbPAR=CALC

Excel AdvancedFilter filters unique data case insensitively, including Cyrillic. And the option corresponding to "Case sensitive" is not there. However, numeric data is distinguished as numbers and strings (?).

Quite obvious to me since the list of items shows only lowercase (Foo, foo = foo) unless I check the sensitive option (only for SF). Would not increase the complexity of the AF but just add a note to the documentation.

No further opinion, recommend to resolve WF (and add it to documentation). The standard filter provides means to run case sensitive and adding the option to the auto filter would clutter the widget further.

(In reply to Mike Kaganski from comment #2)
> Adding Eike to CC

Please reopen in case you disagree.

(In reply to Heiko Tietze from comment #5)
> The standard filter provides means to run case sensitive and adding the
> option to the auto filter would clutter the widget further.
IMHO that's not what is requested.. the original description lacks the mention of "No duplication" that seems to be the actual problem here, which becomes (somewhat) clearer when opening the sample document (which is confusing enough though).

I'd rather expect that when filtering is done case insensitive that also uniqueness (no duplication) is case insensitive. Doing so also wouldn't need any new option in the dialog.
Reopening.

Test: "AAA", "aaa", "BBB", "bbb", "AAA"

Current behavior when sorting "Test" = "not empty" with the options

[x] No duplication & [ ] Case sensitive = "AAA", "aaa", "BBB", "bbb"
[x] No duplication & [x] Case sensitive = "AAA", "aaa", "BBB", "bbb"

Expected behavior:

[x] No duplication & [ ] Case sensitive = "AAA", "BBB"
[x] No duplication & [x] Case sensitive = "AAA", "aaa", "BBB", "bbb"

Created attachment 176687 [details]
for testing

I have attached a simple file in order to test the bug, as mentioned by Heiko.

In order to test: just check No duplication, then Check or Uncheck "Case sensitive": not working.

Still the same results as in comment #7, in:

Version: 25.2.0.1.0+ (X86_64) / LibreOffice Community
Build ID: 468d47bdf44a772e989a0adbeb5bb097e3cb4f31
CPU threads: 16; OS: Linux 6.8; UI render: default; VCL: gtk3
Locale: hu-HU (hu_HU.UTF-8); UI: en-US
Calc: threaded