[Koha-zebra] MARCXML to ISO2079

Tümer Garip tgarip at neu.edu.tr
Thu Jan 19 15:06:52 CET 2006



Hi again,
Thanks for the quick response. Let me try and clear some more on the
problems:

Sebastian Hammer wrote:
>Yes, but best if you use the grs.marcxml input record type,which
>represents the record *internally* in a MARCXML-based structure. This 
>will work with the marc21.abs indexing rules in the Zebra distro, but 
>not the older, clunky usmarc.abs rules.

 Using grs.marcxml record type either by zebraidx from the command line
or using ZOOM it is NOT possible to feed Zebra with xml records.
Zebraidx gives the error "MARC record length<25" and does not update.
ZOOM gives encoding error and does not update. So I use recordtype
grs.xml with marc21.abs (had to rename it as collection.abs though). Now
both zebraidx and ZOOM update the zebrasrv.

>This sounds like something is going wrong. When you've loaded MARCXML,
>it should be trivial to get back MARC.

>Have you tried using the YAZ client to get records out? What do they
>look like?

Neither yaz-client nor ZOOM give me MARC. Leader only and lots of space.
They both give the MARCXML perfectly.

>When I use Zebra as the dedicated search engine in an application I
>usually use this technique to save the extra PDU exchange.. you simply 
>add sort directive on the query by ORing terms on as above.. for
instance:

>f @or water @attr 1=4 @attr 7=1 0

This worked with yaz-client but gave "invalid query" with ZOOM. I'll try
and work on this a bit more.

Thanks
Tumer


-----Original Message-----
From: Sebastian Hammer [mailto:quinn at indexdata.com] 
Sent: Thursday, January 19, 2006 3:49 AM
To: Tümer Garip
Cc: zebralist at lists.indexdata.dk; koha-zebra at nongnu.org
Subject: Re: [Koha-zebra] MARCXML to ISO2079


Tümer Garip wrote:

>Hi,
>
>Three questions:
>
>1- Feeding zebra with ISO2079 records I can get back both ISO2079 and
>MARCXML from zebra very well.
>  
>
Yes, but best if you use the grs.marcxml input record type,which 
represents the record *internally* in a MARCXML-based structure. This 
will work with the marc21.abs indexing rules in the Zebra distro, but 
not the older, clunky usmarc.abs rules.

>However since ZOOM supports only XML update I had to change the
>recordType to grs.xml and use MARC21.abs
>
>
>The problem now is that I can get the full MARCXML but only the leader
>part of IS02079 and lots of space. Ofcourse I can change the MARCXML to

>ISO2079 through other modules but I wonder whether I am doing something

>wrong here.
>  
>
This sounds like something is going wrong. When you've loaded MARCXML, 
it should be trivial to get back MARC.

Have you tried using the YAZ client to get records out? What do they 
look like?

Z> find something
Z> format usmarc
Z> show 1

>2- My marc21.abs contains melm 245$a title:w,title:s
>
>when I use $rs->sort(“yaz” 1=4 i>) I get succes in sort but
>$rs->record(0) still gives me the unsorted list?
>  
>
Mike? Is there some trickery to how you use sort from ZOOM?

Zebra has a shortcut that allows you to attach sort criteria directly to

the query. From the CHANGELOG:

>Added support for sort specifications in RPN queries. Type 7 specifies
>'sort' where value 1=ascending, value 2=descending. The use attribute 
>specifies the field criteria as usual.  The term specifies priority 
>where 0=first, 1=second, ...
>
When I use Zebra as the dedicated search engine in an application I 
usually use this technique to save the extra PDU exchange.. you simply 
add sort directive on the query by ORing terms on as above.. for
instance:

f @or water @attr 1=4 @attr 7=1 0

Produces a result set holding the records matching 'water', sorted in 
increasing order by title.

>3- When indexing I want to be able to index apostrophe character to
>NULL rather than SPACE. Character table .chr explains how to set it to 
>SPACE but not to NULL.  This does not help with my indexing when a user

>misses to use the apostrophe in the search. Any suggestions?
>  
>
This is a feature I have been missing myself for years.. at the moment, 
to my knowledge, there is no way to do this...

--Sebastian

-- 
Sebastian Hammer, Index Data
quinn at indexdata.com   www.indexdata.com
Ph: (603) 209-6853








More information about the Koha-zebra mailing list