[Zebralist] A few questions about zebra

Marc Cromme marc at indexdata.dk
Thu Nov 1 09:12:14 CET 2007


Merijn van den Kroonenberg wrote:
> Hello List,
> 
> I have been playing with zebra and a couple questions (4) came to my mind. 
> Maybe someone can help me with them:
> 
> - When is "storeKeys: 1" needed? I had a testcase with "storeKeys: 0" using 
> general record Id's and I could update with zebraidx, but not delete. I got 
> the following warning:
> [warn] cannot delete file above, storeKeys false (3)
> I would expect, when i can update then i can delete too?
> storeKeys requires quite alot of diskspace so if I do not need it, I want to 
> avoid it.

stored keys are needed for some record-lookup operations, namely 
sorting, replacing and deleting.

You can avoid stored keys on static content - like indexes for CD 
searches where space is premium and content never changes.

> 
> - Is it possible to return only the record ID when presenting results after 
> a search?

Yes. you can always display only the internal record number by using
teh special elementset 'zebra::mets::sysno' using SUTRS or XML record 
format. See

http://www.indexdata.com/zebra/doc/special-retrieval.tkl

> Basically I do not want to store the record data in Zebra (or in the 
> original import location). As a search result i only need the record ID, 
> because I present the data from my own source data.

Then you should just write

storeData: 0

in your zebraconfig, and you will not get stored internal record copies.

  http://www.indexdata.com/zebra/doc/locating-records.tkl


> As a workaround I figure I can replace the indexed file by a smaller one 
> with only the record Id in it, but I wonder if there is a better way.
> Maybe there is a way to tell zebra about an alternative way to retrieve the 
> source record? This would be especially usefull when indexing with extended 
> service Insertrecord.
> 
> - Is it possible to tell zebra to store its index/dictionary data separately 
> from the record/key data?
> The idea behind this question is to store only the indexes (and everything 
> needed for the searching) on fast (expensive) storage and the other data on 
> slower (cheaper) storage.
> Not sure if its possible or even relevant, but just something I wondered.


I do not think so.

> 
> - why is bib1 attribute 1032 'Identifier-document' named 'Doc-id' in the 
> bib1.att?
> Or am i missing something here, i am new to the attribute sets stuff.
> 

Very good question. Historic reasons, probably. You can always rename 
any bib1.att index name, as far as you make sure that you use the same 
name in your ABS or DOMXML indexing rules.

Your's Marc Cromme, Index Data

> Greetings,
> Merijn van den Kroonenberg
> 
> Web2All B.V.
> Telefoon: +31 475 775511
> Fax:        +31 475 338290
> Email:      merijn at web2all.nl
> Website:  http://www.web2all.nl
> 
> 
> _______________________________________________
> Zebralist mailing list
> Zebralist at lists.indexdata.dk
> http://lists.indexdata.dk/cgi-bin/mailman/listinfo/zebralist
> 


-- 

Marc Cromme
M.Sc and Ph.D in Mathematical Modelling and Computation
Senior Developer, Project Manager

Index Data Aps
Købmagergade 43, 2
1150 Copenhagen K.
Denmark

tel: +45 3341 0100
fax: +45 3341 0101

http://www.indexdata.com

INDEX DATA Means Business
for Open Source and Open Standards







More information about the Zebralist mailing list