Clyde wrote: >Thanks, the problem was with my test data being incomplete and I was expecting >SYSTEM to have ALL uppercase data to come before ALL lowercase data (similar >to the default sqlite collation), and that SYSTEMNOCASE should be the only one >to group differnet case together. Glad you were able to sort this out! >It would appear that the collation SYSTEM gives *nearly* equal "weight" to >upper and lowcase letters that are the same, rather than a big gap between say >a-A > >How can I get a table or list of what each character "weight" is when the >Collation SYSTEM is used? I don't know. The actual comparison is done by the Windows.CompareString... function and its actual implementation depends on the user/system langauage setting (German might sort different than French). You would have to ask Microsoft for the actual comparison tables or implement your own string comparison function to know 100% for sure what is going on. Ralf _______________________________________________ Delphi Inspiration mailing list yunqa@xxxxxxxxxxxxx //www.freelists.org/list/yunqa