Re: ODBC performance urban legend - was RE: Output from PF to ExcelXLS file in IFS

On 7 Feb 2003 at 21:48, Vern Hamberg wrote:

> ODBC per se, is not the source of poor performacne in ODBC
> applications. Many other things can be the cause, including using the
> base JET database engine in VB, instead of RDO. Many products that use
> ODBC create ad hoc SQL statements, without index support, that simply
> CANnot perform well.
> 
> Excellent performance with ODBC is possible - John Sears has a good
> presentation on this.
> 
> At the same time, I agree that ADO will usually do better. But it,
> too, will suffer from poor index support.
> 

Particularly if you open tables with CMDTBLDIRECT and exploit indexes.  Using 
seek instead of SQL/WHERE when possible can speed things up dramatically.  

If performance is really an issue, use sockets, as pointed out in Scott 
Klement's 
tutorial at 

http://klement.dstorm.net/rpg/socktut/index.html

This is the fastest thing yet...  We use it to bring large volumes of data into 
an 
analysis tool that produces spreadsheets analyzing inventory movement and 
predicts 
trends.