"paddlefoot" <tfa@bellsouth.net> skrev i meddelandet
news:3aafaf8b$1@news.devx.com...
>
> Sorry Arthur, but your way off. When ADO opens a recordset from any
> database, that recordset is completely independent of the index that
> the database builds. In fact, it is not even used. Here is the low down
> on ADO and why it is so in-efficient for large databases:
>
> 1. IT STORES RECORDSETS IN MEMORY!
> - have you ever noticed how it takes so long to open a recordset using
> ado when you have a lot of records. That's right....these records
>
> are being read into ram. How crummy is that! Say you had 128 MB of
>
> ram, and a 200MB access file. (200MB inventory files are not
> un-common) If you tried to open this with ADO, your machine would
> use up all of its available ram, including Virtual Ram on the disk.
>
> Not to mention you would waiting thirty minutes for the file to

open.
> VOMET on ADO!
>
> That in itself is enogh for me to use some other data access technology.
>


You can always use a server side cursor, then it won't copy all the rows to
ram.