[Koha-zebra] Behind the Scenes Updating

[Sebastian Hammer]

Joshua Ferraro wrote:

>Hi again,
>
>Just wondering if the Index Data guys could clarify a bit what's
>going on behind the scenes when Perl-ZOOM updates a Zebra database
>using the new code. Based on my reading of the Zebra docs, it's
>necessary to run 'zebraidx update' on the whole database every
>time a change is made with file record IDs (which it looks like
>we'll be using for Koha). I'm assuming that there's some internal
>function call that the extended profile supports that goes beyond
>the basic command-line functionality, making it possible to update
>a small segment of the database ... am I correct?
>  
>
Hi Joshua,

I don't think you should be using file record IDs when you update 
records  using Z39.50. it's true that when you use file record IDs from 
the command line, one way of doing so is to run the zebraidx command on 
a whole directory, and it will compare the time stamp of each record 
against what it holds internally and update only those that have been 
changed..

However, when you are updating the database remotely, you are explicitly 
feeding Zebra with the records that need to be inserted or modified. 
File record ids don't work because no files are involved in the transaction.

--Sebastian

>Cheers,
>  
>

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