Re: [Koha-zebra] Bad match criteria errors

[Joshua Ferraro]

On Thu, Mar 23, 2006 at 11:59:15PM +0200, Tümer Garip wrote:
> Well unfortunately I had the same. As soon as zebra finds a record
> without an ID it stops.
> Infact that made me look into our marc records. And to my surprise I
> found a considerable amount with no 090$c
> field where biblionumber is supposed to sit.
Could you tell us what SQL you used to find that? I'm running the following
queries:

mysql> SELECT m1.bibid, m2.bibid FROM marc_biblio AS m1 LEFT JOIN 
marc_subfield_table AS m2 ON m1.bibid = m2.bibid AND m2.subfieldcode = 'c' 
WHERE m2.bibid IS NULL;
+--------+-------+
| bibid  | bibid |
+--------+-------+
|   1654 |  NULL |
| 158070 |  NULL |
+--------+-------+
2 rows in set (38.46 sec)

mysql> SELECT m1.bibid, m2.bibid FROM marc_biblio AS m1 LEFT JOIN 
marc_subfield_table AS m2 ON m1.bibid = m2.bibid AND m2.subfieldcode = 'c' 
WHERE m2.bibid=''; 
Empty set (0.00 sec)

(I can live with two records that don't have 090$c)

> I had to write a script that actually went and searched the whole
> database and put the 090$c$d fields in.
> At first I thought this was me probably messing too much with KOHA. But
> if you are using the biblionumber as your ID as well  then this could be
> a general bug that we have to look in KOHA. Is it dropping the
> biblionumber field at some dubious point? Well I hope not.
Could you commit this script to rel_2_2 or email it to one of us so
we can include it in the distribution?

> And some more news on ZEBRA dev. The damn windows platform manages to
> crash the  server even without shadow files as well so I'll have to wait
> and see if Adam can do something about that. In the meantime I have
> started working on some searches as well. You can test them at
> http://library.neu.edu.tr/opac-search.pl, it can be improved.
I get a 404 on that page.

> For the configuration files I decided not to set the apostrophe to space
> in .chr file. Because we have lots of words with apostrophes and lots of
> people just dont put them in. So say you have Alice's and you write
> Alices Zebra keeps it as Alice s and cannot find it. So for keyword
> searching I use 5=102 and that way I can find them either way while I
> kept the phrase search more strict. Ofcourse this finds lots of other
> things but may be we can reduce that by rankings which I have not dealt
> with.
> 
> In the record.abs I send you you probably realized I had to use some of
> gils.att as well. I mainly used those for authoryty indexes but I cannot
> think how it will be possible to have access point anywehere in the marc
> to put the authority id. Mine now are fixed for 100 6.. And 7.. fields
> for two different authority types. And I think it might be very
> difficult to imlement something very global then local on this issue.
It sounds like you've done quite a bit of customization of the Koha 
base, would you consider contributing your solutions by joining 
Savannah and committing your changes? I fear that your Koha will branch
too far from the project and eventually we'll lose the ability to
easily share code.

Cheers,

-- 
Joshua Ferraro               VENDOR SERVICES FOR OPEN-SOURCE SOFTWARE
President, Technology       migration, training, maintenance, support
LibLime                                Featuring Koha Open-Source ILS
address@hidden |Full Demos at http://liblime.com/koha |1(888)KohaILS