Bulk import problems in 3.0 alpha
Hello, I have problems in importing data with bulkmarkimport and fail to see were they actually come from. The file notices.marc (see http://erratique.ch/tmp/notices.marc) contains 12 record in utf-8 encoded UNIMARC. That is characters 26-29 of 100 $a are set to "50##" (where # is space) which should mean utf-8 encoded AFAIK. Here's what I get when I try to bulk import it :
./bulkmarcimport.pl -c UNIMARC -d -file notices.marc deleting biblios .attempting to use non-existent class sorting routine ....no mapping found for [0x96] at position 2 in BÖSIGER g0=ASCII_DEFAULT g1=EXTENDED_LATIN at /opt/local/lib/perl5/site_perl/ 5.8.8/MARC/Charset.pm line 210. ERROR: failed to perform character conversion for record 5 .no mapping found for [0x85] at position 1 in Åren utan nåd g0=ASCII_DEFAULT g1=EXTENDED_LATIN at /opt/local/lib/perl5/site_perl/ 5.8.8/MARC/Charset.pm line 210. ERROR: failed to perform character conversion for record 6 ...... 12 MARC records done in 2.88230204582214 seconds
The thing is that I don't know if this is a problem of my records (though I know of one other ILS that imports them fine), of Koha, of MARC::Charset or of my perl installation. Besides when I do a rebuild_nozebra I get the following errors :
./rebuild_nozebra.pl
***** building BIBLIO indexes ***** *********************************** 2Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182. 3Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182. 4Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182. 6Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182. 8Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182. 10Cannot decode string with wide characters at /opt/local/lib/ perl5/5.8.8/darwin-2level/Encode.pm line 182.
Inserting records... 159 biblios done
*********************************** ***** building AUTHORITIES indexes ***** ***********************************
Inserting...
authorities done
Can anyone provide a hint ? Another thing is that I find it quite hard to actually see the data I import besides trying to search it. Is there a way to get lists of the data I have (say browse my catalog by title), know the number of records I have, browse my thesaurus and authors alphabetically ? Thanks for your answers, Daniel
I have problems in importing data with bulkmarkimport and fail to see were they actually come from. The file notices.marc (see
Are you sure using XML::LibXML SAX parser? Verify with this script: <koha-root>/misc/sax_parser_print.pl which should display something like XML::LibXML..... See details in <koha-root>/INSTALL.debian -- Frédéric
Le 8 janv. 08 à 08:48, Frédéric DEMIANS a écrit :
I have problems in importing data with bulkmarkimport and fail to see were they actually come from. The file notices.marc (see
Are you sure using XML::LibXML SAX parser? Verify with this script:
<koha-root>/misc/sax_parser_print.pl
which should display something like XML::LibXML.....
Thanks, I was not. This solves the problem for rebuild_nozebra.pl however my problem with bulkmarcimport.pl remains. Daniel
In a private message Henri-Damien Laurent suggested to me that the problem may lie in marcflavour not being set to unimarc in koha's system preferences. This was not the case. However he correctly identified that a (wrong) translation from marc8 to utf-8 was occuring and here, he was right. Invoking bulkmarcimport with -s, described as :
s : skip automatic conversion of MARC-8 to UTF-8. This option is provided for debugging.
solved my problem. Thanks for your answers. Daniel
participants (2)
-
Bünzli Daniel -
Frédéric DEMIANS