Re: SQL order by question

[Mark Lazarus <mlazarus@xxxxxxxx>]

Dean,

At 01:58 PM 1/21/99 EST, you wrote:
>Gary,
>
>In a message dated 1/21/99 9:32:00 AM Eastern Standard Time,
>Gary_Lehman@mail.mchcp.org writes:
>
>> Forgive me for sounding sort of ignorant, but when you say build the access
>>  paths do you just mean rebuilding the logicals?  If so, if you build them
>in
>>  order of least used to used the manager knows this?
>
>Yes, the manager knows in what order the logicals were built and searches the
>first n (someone posted the new figure a while back, I don't recall) of them
>in date order before deciding to build its own.  This brings up an
interesting
>question that perhaps someone listening in at IBM could answer.  Why?  

>With access path evaluation taking so little time in comparison to building a
>new access path over a large file, why not remove the access path
>consideration restriction or at least provide a system value?  I too am a
>proponent of specifying the physical file in an SQL statement, but this can
>cause problems in tuning should you recompile a logical for performance and
>then have the aforementioned n logicals recompiled for a different reason.
>Programs that once ran well suddenly start dying, and nobody knows why
because
>the guy/girl that recompiled the original logical for performance didn't
>document it adequately and has since taken that $30K signing bonus to work
>somewhere else.  With the advent of data warehousing, client/server, and
>terabyte storage capabilities, why do we still have this antiquated access
>path evaluation restriction?

 There is a parm on OPNQRYF to force the optimizer to evaluate all access
paths.

 -mark

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---