Openrefine: Text facet sort by name should use case & diacritic insensitive collation
_Original author: tfmorris (November 12, 2011 19:51:37)_
Currently lowercase characters sort after all upper case characters so 'T' and 't' are in wildly different spots and international characters collate at the very end so that 'Österreichische' is miles from the 'O's.
We should fold both case and diacritics to their base forms.
_Original issue: http://code.google.com/p/google-refine/issues/detail?id=482_
tfmorris
All 3 comments
_From tfmorris on November 12, 2011 20:33:31:_
r2371 makes the sorting order case insensitive, but Javascript doesn't appear to have a built-in diacritic folding method, so that'll be a little more work.
After I committed the "fix" I discovered that this may actually be a browser-specific bug/difference, but it doesn't appear that there's been much progress in fixing it, so we probably should assume that the current state is going to exist for a while.
http://code.google.com/p/v8/issues/detail?id=459
There's a code snippet here that can be used to scrub diacritics: http://lehelk.com/2011/05/06/script-to-remove-diacritics/
tfmorris
on 15 Oct 2012
Waiting 8 years has its advantages - there's now ECMAscript support for Intl.Collator which collates letter case and diacritic forms together (according to locale specific rules).
tfmorris
on 7 Jul 2020
The default localeCompare() implementation collates diacritics together, at least for the en-US locale with Chrome, but presumably collates things the way users expect in all locales, so I think we can close this.
tfmorris
on 31 Jul 2020
Related issues
kushthedude
·
3Comments
katrinleinweber
·
3Comments
thadguidry
·
4Comments
asyrul21
·
3Comments
wetneb
·
3Comments
Most helpful comment
Waiting 8 years has its advantages - there's now ECMAscript support for Intl.Collator which collates letter case and diacritic forms together (according to locale specific rules).