Re: embedded SQL

Richard J Woodland wrote:

> <snip>

> My preferred method, at least for the moment, is to use the first
> method. My cgi script is perl, it produces SQL queries in temporary
> files, uses the 'system' function to call sql*plus with the '@' to get
>
> the script from the temporary file. The script contains a SPOOL
> command
> to send the output to another temp file, which is read by the perl
> script, parsed and converted to an HTML response page. Sounds like a
> pain, but it's actually quite straighforward, in most cases.

This technique is >ok< however it does not scale for really large connection/request rates on the web server. There is too much latency introduced with all the disk and cpu activity associated wih forking a sqlplus session and creating a session. Also, for very large tables you may want the granularity of the perl DBI interface to step through a cursor and perform an update all within a very small while loop.

-Lane

<more snipped in the interest of ecology>

Lane Sharman
Pour chasser le chagrin,
il faut du vin. Received on Wed Jun 25 1997 - 00:00:00 CDT