[Zebralist] ICU and Truncation

Adam Dickmeiss adam at indexdata.dk
Wed Feb 20 09:24:14 CET 2008


Thien Ho wrote:
> Hello,
> 
> I attached a tar ball including my configuration files, sample
> records. You may need to change modulePath in zebra.cfg to your module
> directory. The included modules is for RHEL4.
> 
> Problem:
> I have a problem with truncation when enable ICU index. I'm using
> Zebra 2.0.26, YAZ 3.0.22 on CentOS 4.
> 
> Step to produce problem:
>  - Extract the tar ball.
>  - Index records
>  - Start Zebrasrv
>  - Search using yaz-client.
> 
> tar jxpf zebra_icu.tbz'. A directory called zebra_icu will be created.
> cd zebra_icu
> zebraidx init
> zebraidx update record/
> zebrasrv
> yaz-client @:9999
> 
> Z> f @attr 1=1016 @attr 5=102 .*
> The above search returns 3 records, which is fine.
> 
> 
> In my configuration, I map attribute value 5000 to field 852,
> sub-field p. So searching using attribute value 5000 should returns 3
> records, but it only give me 2 records.
> 
> Z> f @attr 1=5000 @attr 5=102 .*
> This one returns 2 records only.
> 
Thanks for your data. We will see if we can sort out the issues with ICU 
  and truncation. In any case, it's quite different from the .chr-way 
(old system), since Zebra is working on the ICU sort-normalized strings 
and we don't know if it makes sense to truncate these (at all).

The problem you are seeing is probably identical to Zebra bug #2049.
http://bugzilla.indexdata.dk/show_bug.cgi?id=2049

/ Adam

> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Zebralist mailing list
> Zebralist at lists.indexdata.dk
> http://lists.indexdata.dk/cgi-bin/mailman/listinfo/zebralist




More information about the Zebralist mailing list