>>Can anyone help me with the right syntax for an indexkey that sort the following data:

>>cAuktCID  cObjectCID  nOffer    CID
>>"    21"  "    83"    567.00  "   129"
>>"    21"  "    83"    567.00  "   126"
>>"    21"  "    83"    567.00  "   128"
>>"    21"  "    83"    650.00  "   125"
>>
>>
>>I want it to look like this when it's indexed.
>>
>>  Desc      Desc       Desc     Asce
>>"    21"  "    83"    650.00  "   125"
>>"    21"  "    83"    567.00  "   126"
>>"    21"  "    83"    567.00  "   128"
>>"    21"  "    83"    567.00  "   129"

>>I have tried with this, but i does not work:
>>cAuktCID + cObjectCID + STR(nOffer) + STR(VAL(cid)*-1) DESCENDING
>>
>>The 3 first columns gets sorted right but CID-column always
>>ends up in descending order although i use -1 in the calculation.
>>What am i missing here ?
>
>How about

cAuktCID + cObjectCID + STR(nOffer) + STR(999999 - VAL(cid)) DESCENDING

>Also, when indexing on a character expression, remember that VFP uses fixed length index keys. Unless you force the length of the key in the expression, its length will be based on the first record in the table, with unpredictable results. As it happens, in your case STR() defaults to (forces a) length of 10 since the length is not explicitly specified, but this is probably unnecessarily long, making your .CDX files larger and therefore less efficient than necessary.

Hi Al

This is the solution I was trying to achieve. This one I will use. It is nearly the same as Daniel Gramunt mention in another reply.

Thank's Al ,Craig and Daniel for helping me with this one.

/Torgny