Charles Jardine wrote:

> 1. ora_max_nested_cursors is not relevant. This value is used
>   only in the case where a result set has columns of type REF CURSOR,
>   resulting in nested fetches. If you don't have a 'CURSOR( ... )'
>   expression in the select list of a 'SELECT' statement, you don't
>   have nested cursors.

OK, that makes sense to me -- they didn't seem intuitivel "nested".
As you can tell, I was just trying anything that seemed close.

Unfortunately, removing the setting of ora_max_nested_cursors 
doesn't make it go faster either... :)

> 2. Does you result set have LONG or LONG RAW columns? Caching is
>  disabled if you have these.

I checked, and they're all VARCHARs and one NUMBER type.  Interesting
observation though -- I'd never heard that before.

Thank you for your reply.  Anything else you can think of for me to 
check?  

If you do a similar thing in your environment, do you get
the same behavior, or does DBI cache ahead when reading from a
cursor returned from a stored proc?

Just in case, I replaced the stored proc call with the SQL from
within the stored proc, and sure enough the sample DBI/DBD::Oracle 
code caches ahead very nicely.  So this problem seems limited to 
pulling results from a cursor. 


Thanks again!

Joel Noble
[EMAIL PROTECTED]




Reply via email to