Cetin,

It is actually the SYS(15) I'm looking at. I can create a compatible index with Clipper using a Nor() function in both VFP and Clipper. (The problem is Norwegian characters ÆØÅ and æøå that is supposed to be sorted after XYZ and that I here are giving the values [\]. In Fox 2.x the collate alternatives are wrong.) It would be easy to change the indexes in the old Clipper program with the same function. (The Clipper program do use CDX and Fox record locking.)

USE MyTable
SET COLLATE TO "machine"
INDEX ON Nor(NAME) TAG NAME
brow

FUNCTION Nor(x)
  LOCAL cRet
  cRet=SYS(15,SPACE(32)+SPACE(15)+;
  '0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]'+;
  '...ABCDEFGHIJKLMNOPQRSTUVWXYZ{|}..CYEA\A]CE[EIII[]E[[O\OUUY\Y'+;
  '\.\..AIOUNN...............AAA.....\\.......AA.........DEEEiII'+;
  'I.....I.OBOOOO÷÷÷UUUYY÷÷÷÷÷÷÷÷÷÷÷÷......',X)
RETURN cRet

You suggest I can create an indexed view in a DBC. That is fine for reading. But what about updating the 2.6 table? (From .Net)


>
>Einar,
>I wouldn't index a FP2.x table shared with clipper. You might render it useless. Instead create a temp (or permanent) dbc and view in it. You could index the view.
>PS: For a collation sequence IMHO it's better to use char mapping (ie: sys(15) - hope remembering right).
>Cetin