[Koha-devel] Imported MARC notice and Authorities: : inconsistent database

[Frère Sébastien Marie]

Hello,

As we obtain no reply for an inconsistent database, I will explain again the problem, with SQL for examples.

Currently, after using imported set of notices and using it we have the following results:

SELECT count(*) 
 FROM biblioitems
 WHERE extractvalue(biblioitems.marcxml, '/record/datafield[@tag="700"]/subfield[@code="9"]') != ''
   AND extractvalue(biblioitems.marcxml, '/record/datafield[@tag="700"]/subfield[@code="9"]') NOT IN (
    SELECT authid FROM auth_header );
+----------+
| count(*) |
+----------+
|      121 |
+----------+

english-version: If I extract from MarcXML (on table biblioitems) the 700.9 (Primary Author, koha_internal_code) and search in auth_header, we currently have 121 items with no-existent 700.9 .

This is only a part of the problem: some 700.9 are existent but wrong (points to bad authority).

SELECT count(*)
 FROM biblioitems, auth_header
 WHERE extractvalue(biblioitems.marcxml, '/record/datafield[@tag="700"]/subfield[@code="9"]') = auth_header.authid
   AND extractvalue(biblioitems.marcxml, '/record/datafield[@tag="700"]/subfield[@code="a"]') != extractvalue(auth_header.marcxml, '/record/datafield[@tag="200"]/subfield[@code="a"]');
+----------+
| count(*) |
+----------+
|        3 |
+----------+

Any suggestions ? This seems a bug (in import of notices where old koha_internal_code is not removed), but I would prefered have your opinion before submit a bugreport.

And I will like some clues to correct the inconsitences !

Thanks.

-- 
Frère Sébastien Marie
Abbaye Notre Dame de La Trappe
61380 Soligny-la-Trappe
Tél: 02.33.84.17.00
Fax: 02.33.34.98.57
Web: http://www.latrappe.fr/

On Sat, Jan 29, 2011 at 12:13:28PM +0100, Frère Sébastien Marie wrote:
> Hello,
> 
> Recently, we have imported a set of notices, from another abbey, in
> the reservoir.
> 
> When we create a new examplar using the previous imported notices,
> the field 700.9 (authority - internal koha number) refering the
> auth_id isn't updated or removed (still reference the old authority
> number, in the other catalog).
> 
> As result, we have an examplar with 700.a (authority name) correct,
> but with an internal reference wrong (not valid in us catalog).
> 
> If we search all notices linked with an authority, there is wrong
> attribution.
> 
> Next an example.
> 
> Starting with:
>  - Notices in reservoir:
>    - Title_A / Author_A (auth_id = 1) [from import]
> 
>  - Authority in catalog:
>    - Author_B (auth_id = 1) [created locally]
> 
>  - Notice in catalog:
>    - Title_B / Author_B (auth_id = 1)
> 
> 
> Next, we create a new examplar based on notice in reservoir, result:
>  - Authority in catalog:
>    - Author_B (auth_id = 1) [created locally]
> 
>  - Notice in catalog:
>    - Title_A / Author_A (auth_id = 1)
>    - Title_B / Author_B (auth_id = 1)
> 
> (doubt if Author_A added in authorities or not)
> 
> Now, if search notice attached with Author_B (auth_id = 1):
>  results:
>   - Title_A / Author_A (auth_id = 1)
>   - Title_B / Author_B (auth_id = 1)
> 
> The database is inconsistent.
> 
> Where is the problem:
>  - a misused of imported MARC: an option somewhere should be set...
>  - a bug in importing MARC: 700.9 should be cleaned before import...
>  - another possibility ?
> 
> Currently, no bug report filled, as I don't no if it is a misusing
> or a bug.
> 
> Thanks.
> -- 
> Frère Sébastien Marie
> Abbaye Notre Dame de La Trappe
> 61380 Soligny-la-Trappe
> Tél: 02.33.84.17.00
> Fax: 02.33.34.98.57
> Web: http://www.latrappe.fr/