Still getting err"Putstrarry fail for key user@.."

[John Jaberi]

Hello,

I had problem with sizedb indexing and with a couple suggestion of
changing the multiplier from 2 to 4, it has not change my
problem.
I have modified the sizedb prel script to have bigger multiplier
which it should fix my probel but it hasn't.
I have more description on this probelm down below.
I would be apperciate for any comment.

********Frist time question put to the ph mainling list
am running qi-2.4 and it was working fine until add
> some more data to the database.
> When I try to rebuild database using ph-build.sh shell
> script, then I'll get these messages:
>
> ****************
> :sh ph-build.sh
> PH database build script started...
> Making a working copy of prod.cnf for building the new database...
> Calculating size...
> Executing credb...
> /usr/local/qi-2.4/util/credb: creating database prod-new
> prod-new.idx : 16000 bytes
prod-new.idx : 16000 bytes
> prod-new.dir : 400 bytes
> Executing maked...
> maked: building database prod-new
>
> Executing makei...
> makei: indexing database prod-new
> nents = 478
> next_id = 478
> 1 to sort
> sent indicies for 478 dir entries to sort.
> from sort.
>
> **********I can not query these users from ph server****
> **********it will not find any match******
> putstrarry failed for key fja...@metro2.k1 (1 elements)
> putstrarry failed for key fletcher@metro2. (1 elements)
> putstrarry failed for key ge...@metro2.k12 (1 elements)
> putstrarry failed for key ghanka (1 elements)
> putstrarry failed for key gibbs (1 elements)
putstrarry failed for key gibbs (1 elements)
> putstrarry failed for key gking (1 elements)
> putstrarry failed for key glenda (1 elements)
> putstrarry failed for key gmartins (1 elements)
> putstrarry failed for key zimmer (2 elements)
> 2706 from sort
> indexed 1615 unique strings out of 2706 total.
> Executing build...
> build: building sequence file for database prod-new
> 1 () to sort.
> 500 to sort
> 1 from sort.
> 499 from sort
> Tree built, 15 nodes
> 16 items (12 bytes each) written on file prod-new.bdx
> Moving database into place...
> PH database build script complete.

did I confuse this with something else ?) which means that you have to
increase the multiplier, the number 2 in sizedb script.

If anyone are interested I could mail the script I have that is a modified
version of ph-build.sh and lets you specify which database you want to build,
which multiplier you want to use, etc. which makes it more general and
flexible. Let me know.

****This is responds got from ph mailing list******

=========================================================================

To un-subscribe from this list, send email to LIST...@POSTOFFICE.CSO.UIUC.EDU with the command

UNSUB INFO-PH

in the *body* of the message.

[P-P...@qualcomm.com]

Try modifying the sizedb function to count the number of occurrences of
each unique word. Then sort by count and print the top 100 or so
winners. It may be that there's some common word like an areacode
that's overfilling the hash table. What to do next is dependent on
what's found.

/pbp