Right now the 260 index look like this:
<!--record.abs line 119: melm 260$a pl:w,pl:p-->
<index_subfields tag="260" subfields="a">
<!--record.abs line 120: melm 260$b Publisher:w,Publisher:p-->
<index_subfields tag="260" subfields="b">
<!--record.abs line 121: melm 260$c copydate,copydate:s-->
<index_subfields tag="260" subfields="c">
<!--record.abs line 122: melm 260 pl-->
The 264 only indexes the entire thing.
<!--Index for RDA 264 field-->
We should break the 264 up for better/more accurate searching.
Split-up the index like 260 was my first intention when I added the new index "provider" also was splitted up in 'Provider-place', and 'Provider-date'.
NOTE: Clarified that the latter two do not exist in the documentation.
If we want to collect all indexes in Provider -this is the true intention of the index, not just collect 260, 264 but also 270, 257- works if you change etc/zebradb/ccl.properties as follow
pv Provider Place-publication Publisher copydate
the exmaple above follows yaz documentation at <http://www.indexdata.com/yaz/doc/tools.html#CCL> in section 220.127.116.11 Qualifier alias.
This at least is not more accurate but will launch more search results
Please let me know which solution do you prefer or/also I can attach two patches to see which is most suitable.
At first glance I thought it was Katrin who file this bug, but I was wrong.
What do you think about comment 1, or could you propose something at this point?
Created attachment 69852 [details] [review]
Bug 15704 - Split up zebra indexing of RDA publisher information
1 - Add a record with a unique publisher "Supercalifragilistic" in the
264 b field
2 - Search for the value
3 - Record not found
4 - Apply patch (may need ot copy the .xml file into koha install)
5 - Reindex all the things
6 - Search for the value
7 - Success!
Using the most basic record with the 'Default framework', I can't reproduce the issue of not finding a record from its publisher.
Here's what I did:
* Since the 264's weren't visible by default in the editor, I went ahead and made them show up by changing the properties of admin/biblio_framework.pl. Otherwise, I had no way of adding content to those fields
* I created a new record, filled the info.
* * I also made sure to fill the other fields of 264 with garbage data, just in case
* Rebuilt zebra
* Searched for the string I added in 264b: supercalifragilisticexpialidocious
* I'm redirected to the new record's page
Am I doing something wrong here? Why is it showing up on my master?
>Am I doing something wrong here? Why is it showing up on my master?
How did you search for it? Since we switched to dom indexing you can always search everything in the record within the keyword index, so it should always show up.
There also was bug 14198.
The interesting question is if you can find it in the specific indexes for place, year and publisher. :)
Not sure this is a bug.
Publisher in the 264$b appears to be searchable via the publisher index even without the patch applied (e.g., search.pl?idx=pb&q=superkapow). Also searchable via a plain keyword search.
Hi Cab! Are you sure the record used in your example search didn't also have that data in some field that *is* indexed as publisher? I just tested this on a library migrating to 18.05.05 and records that have publisher data only in the 264 don't come up in a publisher search.
You may be right, Andrew. Not sure how I was testing back before the fall -- lost in the mists of time -- but I most likely did overlook something ...