Re: Trigger Buffer Question

["Nelson C. Smith" <ncsmith@xxxxxxxx>]

Even before the arrival of blobs, I already have the problem of triggering a
file which has records of 24,000+ bytes.  A before & after version of the
record will not fit in a buffer limited to 32,767 bytes.  I suspect that
data management actually writes out the buffer as long as it needs to be,
but we have (by convention) been artificially limiting it down to 32,767
bytes.  I haven't proven this to myself yet, but if it is indeed true, then
perhaps blobs will be written out as they are and we will have to use
pointer math to find the images in the buffer.

I'll get back to you on my experiments unless someone else already knows the
answer.  Anyone from IBM listening?  Of course first, I have to fill up a
24,000 byte text field with something I can recognize  in the buffer :-)

-----Original Message-----
From: Douglas Handy <dhandy@isgroup.net>
To: MIDRANGE-L@midrange.com <MIDRANGE-L@midrange.com>
Date: Monday, August 02, 1999 12:16 AM
Subject: Re: Trigger Buffer Question


>Nelson,
>
>>The question that comes up is what does data management
>>put into the trigger buffer when the file record is really big?
>
>Until more info on the fall PTF to enhance DB2, I think this is
>unfortunately not yet public knowledge.  The online V4R4 DDS Reference
>still shows the following limits:
>
>"The maximum number of fields in a record format is 8000. If any of
>the fields in the record format are date, time, timestamp, variable
>length, or allows the null value, then the actual maximum number of
>fields can be less than 8000. The maximum number of fields can vary
>depending on the number of fields and combinations of fields that
>occur within the record format. The maximum number of bytes in a
>record format is 32 766 if variable length fields are not included and
>32 740 if variable length fields are included. Table 1 describes rules
>for determining the total length of the record format. "
>
>Presumably something will change with the addition of support for
>blobs, etc.  But I'm not in the know.  I just wanted to point out
>that, until this fall anyway, the array overlay technique used prior
>to pointer arithmetic is capable of handling current limits.
>
>Doug
>+---
>| 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
>+---
>

+---
| 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
+---