OK.
Any Ideas? I have a Type 30 file that if I use the SAMPLE modifier in the
SELECT still returns the entire file,
However, if I use the SAMPLE modifier in a LIST command on that same file, it
only returns the sample amount.
Now, it gets stranger, If I do a SELECT with the SAMPLE modifier, it
George, I checked out system and we have experienced this issue before.
Check if you have built an index on @ID. If you have, this is what is causing
your issue. Using the NO.INDEX keyword will give you correct results. I don't
have an explanation for the behavior. Let us know how it goes.
I'll try to set one up. Can't promise how it will go though, as depending on
how many people bring laptops, it will diverge in to me walking around helping
people out more than me displaying on my screen.
Cheers,
Dan
-Original Message-
From: u2-users-boun...@listserver.u2ug.org
George,
Does the file have an index on the id?
If so, try adding the NO.INDEX keyword.
I've something similar @ UV10.2.10:
--
SELECT PARTS SAMPLE 2
26417 record(s) selected to SELECT list #0.
-
SELECT PARTS SAMPLE 2 NO.INDEX
2 record(s) selected to SELECT list #0.
-
DELETE.INDEX
YUP.
I just isolated it to the index of @ID.
I changed the original filename, recreated a new file
Copied all the data
Copied all the dict
And finally, recreated the index (@ID)
Testing the sampling on each step, all went well until I recreated the index,
then it bombed again.
And yes, using
Oka work around that I didn't think would work...but did...
1. COPY VOC SAMPLE to SAMPLE.ORIG
2. CREATE A VOC SAMPLE ENTRY as below
ED VOC SAMPLE
2 lines long.
: P
0001: PH
0002: NO.INDEX SAMPLE.ORIG
Bottom at line 2.
: EX
ED VOC SAMPLE.ORIG
3 lines long.
: P
0001: K
0002: 22
Maybe I am missing something, but why would you want an index on @ID?
JRI
-Original Message-
From: u2-users-boun...@listserver.u2ug.org
[mailto:u2-users-boun...@listserver.u2ug.org] On Behalf Of George Gallen
Sent: Friday, May 24, 2013 10:18 AM
To: U2 Users List
Subject: Re: [U2] UV
In thinking about itI'm not sure, but it seems to increase the selection
speed (considerably). It is a type 30 file,
So possibly the index makes up for a hashing algorithm where the modulo is too
small? And the makeup of
The @ID causing more similar hashing?
-Original Message-
If you are doing it because of performance, I am thinking you need to re-size
the file (as you suggested). This will not just impact SELECTS, but any IO.
Also keep in mind you are impacting performance just by having the index. All
that being said, getting the time to resize is something
For this particular file, it only gets modified once a day - and then being
perfectly tuned, might shave a couple minutes off the process
To updates it.
The Selections are done during the day for searching, and the index helps that
- so for this particular file, it's fine the way it's working.
If your record key is comprised of something meaningful you might want to query
against it. For example if your key is say STATE: '*': CITY and you want all
the records for a given state ... SELECT FILE WITH @ID FL*] .
Perry
-Original Message-
From:
On 24/05/13 15:57, George Gallen wrote:
In thinking about itI'm not sure, but it seems to increase the selection
speed (considerably). It is a type 30 file,
So possibly the index makes up for a hashing algorithm where the modulo is
too small? And the makeup of
The @ID causing more
12 matches
Mail list logo