Hi Bill,

I have never used SDT, so don't know how it re-indexes, but the best way is to DELETE TAG ALL and recreate the indexes to avoid index bloat.


Regards
Doug Johnston



>We are wrapping up an application. During testing we found that when we reindex the tables, the performance drops drastically. Here is the situation:
>
>We have a table with about 32,000 records. The data in this table was generated by importing data from another data source with a VFP import program we wrote. The table has about 40 fields, one being a memo field.
>
>It is accessed in a fairly complicated view with several joins in a run-time query system. To return a single record, including the overhead to draw the form, it took less than 1 second. All is well.
>
>We reindexed the file. After the reindex, the same exact test takes over 50 seconds to return the same single record.
>
>We tried a couple things: We use SDT. We reindex the table using SDT - no difference. We reindex the table using the standard tools in Mere Mortals - no difference. We tried restoring the data and taking single passes through each of the above scenarios. All had the same result: Untouched table = 1 sec, reindexed table = 50 secs.
>
>If we Pack the table, the performance is almost as good as new, about 2 seconds.
>
>Note that either reindexing using SDT or VFP or Packing all created a CDX that was smaller (1.5MB down from 2.4MB)
>
>Looking in SDT or the table modifier in VFP shows no signs of corruption.
>
>Why would reindexing the table make the program run 50 times slower?