Summary: | ctype index Zebra configuration | ||
---|---|---|---|
Product: | Koha | Reporter: | Fridolin Somers <fridolin.somers> |
Component: | Searching | Assignee: | Fridolin Somers <fridolin.somers> |
Status: | Failed QA --- | QA Contact: | Marcel de Rooy <m.de.rooy> |
Severity: | normal | ||
Priority: | P5 - low | CC: | alex.arnaud, jcamins, koha.sekjal, m.de.rooy, paul.poulain, veron |
Version: | Main | ||
Hardware: | All | ||
OS: | All | ||
GIT URL: | Change sponsored?: | --- | |
Patch complexity: | --- | Documentation contact: | |
Documentation submission: | Text to go in the release notes: | ||
Version(s) released in: | Circulation function: | ||
Attachments: |
Proposed patch
Proposed patch (updated) Bug 5295 : ctype index Zebra configuration [Signed-off] Bug 5295 : ctype index Zebra configuration |
Description
Fridolin Somers
2010-10-08 09:37:04 UTC
Created attachment 2684 [details] [review] Proposed patch Created attachment 7614 [details] [review] Proposed patch (updated) Updated patch. Please signoff. Fridolyn, Could you please resubmit this as a git patch? Created attachment 9655 [details] [review] Bug 5295 : ctype index Zebra configuration Patch submitted originally by Fridolyn SOMERS Reformatted as a git patch Tiny patch. Tested searches on Content-type and it works. So signing off Created attachment 10175 [details] [review] [Signed-off] Bug 5295 : ctype index Zebra configuration Patch submitted originally by Fridolyn SOMERS Reformatted as a git patch Signed-off-by: Alex Arnaud <alex.arnaud@biblibre.com> Replaces bib1.att ctype with Content-type, which is the correct index according to external sources, as well as ccl.properties. Marking Passed QA. mmm... greping ctype, I see that there are a lot of references still here: etc/zebradb/ccl.properties:ctype Content-type etc/zebradb/marc_defs/normarc/biblios/record.abs:melm 008 date-entered-on-file:n:range(data,0,5),date-enter ed-on-file:s:range(data,0,5),pubdate:w:range(data,7,4),pubdate:n:range(data,7,4),pubdate:y:range(data,7,4),pubdate :s:range(data,7,4),pl:w:range(data,15,3),ta:w:range(data,22,1),ff8-23:w:range(data,23,1),ff8-29:w:range(data,29,1) ,lf:w:range(data,33,1),bio:w:range(data,34,1),ln:n:range(data,35,3),ctype:w:range(data,24,4),Record-source:w:range (data,39,0) etc/zebradb/marc_defs/marc21/biblios/record.abs:melm 008 date-entered-on-file:n:range(data,0,5),date-entere d-on-file:s:range(data,0,5),pubdate:w:range(data,7,4),pubdate:n:range(data,7,4),pubdate:y:range(data,7,4),pubdate: s:range(data,7,4),pl:w:range(data,15,3),ta:w:range(data,22,1),ff8-23:w:range(data,23,1),ff8-29:w:range(data,29,1), lf:w:range(data,33,1),bio:w:range(data,34,1),ln:w:range(data,35,3),ctype:w:range(data,24,4),Record-source:w:range( data,39,0) COULD SOMEONE USING MARC21 confirm that this patch does not break marc21 ? I've some doubts... I think indeed MARC21 will be broken because it uses ctype in record.abs. By now, "ctype" works for MARC21 but it's a best pratice to use long names in configuration and short names only in CCL queries, no? Can we replace "ctype" by "Content-type" for all record.abs ? (In reply to comment #9) > Can we replace "ctype" by "Content-type" for all record.abs ? Yes, you should, if you rename it in bib1.att. The patch at this moment will raise the warning: [warn] Index 'ctype' not found in attset(s). Nothing will be found anymore with ctype=... Changing status to reflect need for adjustment. Furthermore, you remove "t=b" (right and left truncation). I think that is okay (at least for left truncation). But note that this field is NOT a single letter. Actually, it contains (in MARC21) four characters. But they should be taken as four separate letters. That is currently a problem in Zebra; e.g. if I enter abc in Content-type, I cannot find this record when searching for ctype=b. In the correct interpretation of this marc field, I would like to find it when searching for a, b or c. Probably, this problem is outside the scope of your report. > M. de Rooy : Furthermore, you remove "t=b" (right and left truncation)
I agree with you.
But I think truncation should not be implicit.
You can add truncation in query search.
Or create a second index CCL that contains truncation.
ie : Content-type-truncated 1=1034 t=b
Still valid? |