Hi,
there had been a long discussion of Derby-134 ("Sorted string columns are sorted in a case sensitive way"), which in the end looked to me as being solved, i.e. someone has supplied a "good-looking" patch. Now is there any chance of a binary build coming up in the near future, where this patch would be included?
Talking about "ORDER BY", after reading the manuals about optimization and in particular http://incubator.apache.org/derby/manuals/tuning/perf45.html#Cost-Based+ORDER+BY+Sort+Avoidance
I'm still not sure about a few things. My problem is that I have a GUI with a table widget that displays >1500 rows, and I want to make use of Derby's index structures to have fast column sorting. Maybe that simply won't help here, because I need the ordering to be both locale-sensitive and case-insensitive.
With an ascending index on a VARCHAR column:
1. does that index default to case-insensitive ordering of the column?
2. does it maintain the ordering in a locale-sensitive way, e.g. by using java.text.Collator? (e.g. using the default locale active at index creation time)
3. will "ORDER BY ... DESCENDING" make use of the index by simply reversing its order, or will it result in full re-sorting with performance-penalty?
Thanks for any information, J�rg.
