I suspect the following:

Commands like LOCATE use indexes (they don't need to be active) for optimization.

If an index is corrupted, you may get incorrect results. The more involved expression will override the use of the index.

To test this, either reindex, or temporarily turn off Rushmore Optimization.

HTH,

Hilmar.



>I have run into an anamoly with the LOCATE function which has me baffled.
>
>I have used this funtion for almost 20 years without incident and now I have bumped into something.
>
>I have a table CLUB1 with a field CLUBNO with a structure of N(10) integer.
>
>I have the CLUB1 table in a number of directories in accordance with the Company using the program.
>
>One of the CLUBNO values is 1027. I then set a = 1027.
>
>I then do LOCATE for CLUBNO = a
>
>When I use it in the Main Directory, this function works fine. When I use it in a subdirectory of the Main Direcory the function doesn't work. I have to use: LOCATE for INT(CLUBNO) = a to get it to work properly.
>
>The CLUB1 tables in both directories are identical. No filters and no indexes are active on the files.
>
>I am using the CLUB1 file by putting the exact directory reference in front of it. When I run DISP STAT, the subdirectory shows up correctly.
>
>I have tried it with EXACT ON and EXACT OFF and this makes no difference.
>
>I was so perplexed by this, I ran CHKDSK to see if I had hard drive problems and this didn't help either.
>
>I am using VFP 6.0 under Windows NT 4.0 .
>
>This is absolutely weird. If anyone has come across this, I would really like to hear from you.
>
>Mike Smith