AREV GETLIST NOT WORKING CORRECTLY (General)
At 10 JUN 2004 02:06:43PM dsig@sigafoos.org wrote:
AREV3.1, WIN2K, standalone machine.
Have a getlist then a loop with readnext will get stuck on a 'block' of ids (saved list creates multiple blocks). By stuck I mean that say
LISTNAME
LISTNAME*1
LISTNAME*2
….
as it processes through it might get stuck on the first id of LISTNAME*2.
We started with a list of approx 1.2meg keys. we have reduced the number of keys down to 8k
Currently dealing with list of 8297 keys, 6 blocks with id size of approx 20 +- 10. The Keys are 3 part. There are no MFS's on the file but there are 2 indexes.
Currently the process simply gets the list then loops through the ids, reading in a record and updating 4-5 fields, then writing out. We now write out the id into a test file, at readnext time, to see if we see the same id again. and we do.
we have created a little program simply to getlist, readnext, write to testfile to see if we see the id a second time. We do NOT have the error.
The routine does xlate off of two different files ( bringing in the entire record ). Then parse out the appropriate data and shoves in to the main record. then writes the data out.
There are no external subs (though there are 2 interal. neither do anything that could cause any problems)
Anyone seen anyting like this?
dsig@sigafoos.org.com onmouseover=window.status=the new revelation technology .. a refreshing change;return(true)"
David Tod Sigafoos ~ SigSolutions
At 10 JUN 2004 07:16PM dsig@sigafoos.org wrote:
This turned out to be an odd memory problem (at least that is all we can think of). After several other tests we moved lone environment to another machine and it worked correctly.
Never seen an issue like this .. oddly reproducible on the original machine but at least we now know
dsig@sigafoos.org.com onmouseover=window.status=the new revelation technology .. a refreshing change;return(true)"
David Tod Sigafoos ~ SigSolutions
At 11 JUN 2004 06:25PM Steve Smith wrote:
You weren't holding your mouth the correct way.
At 12 JUN 2004 12:14PM dsig@sigafoos.org wrote:
Don't think my mouth was involved but I know that IT's memory was as it used to be.
Also found that the clients LIVE INDEXER was running on a pII win98 with 16meg. it also seems to have memory issues. funny how LIVE boxes can be less that the developers And they wonder about there speed issues.
dsig@sigafoos.org.com onmouseover=window.status=the new revelation technology .. a refreshing change;return(true)"
David Tod Sigafoos ~ SigSolutions
At 13 JUN 2004 08:46PM Steve Smith wrote:
You might want to run an expendable version of CALCULATEX