Walter,

It's possible that the size of the structural CDX is a factor. You don't have control of how Fox handles the CDX, but I have seen where deleting some of the CDX tags (moving them to separate IDX files, or just doing away with them) has increased file-open performance over a slow WAN.

Just a SWAG, but it might be worth trying.




>Hi Mike,
>
>After some experimenting on a WAN network I discovered that opening lot's of tables resulted in very poor performance. It seemed to be that small tables ( smaller than a few kb) loaded faster (with either the USE statement as loading with the dataenvironment) than longer tables.
>
>On my 128 kb WAN some tables took about one or two seconds to open and other larger tables took about 8 to 11 seconds to open. I concluded that VFP buffers a portion of the table when opening the table which is slowing things down.
>
>I also came to the conclusion that many of the poor performance of form instantiation cases might be due this buffering mechanism.
>
>No my question is, is there any way to have control over the amount of data VFP buffers when opening the table? If not I'd like to adress this point as a ER. I really think that such an mechanism could optimize many performance problems (on a WAN and even over a LAN network) when opening tables.
>
>Regards,
>Walter,