TREX and foreign characters

[Former Member]

Hi,

When indexing a document with the swedish word "älg" in it you can find it by searching for "älg" (of course). The word is also indexed in the form of "aelg" for users without the possibility to use the ä-character.

Is it possible to change this behaviour so that the word also is indexed in the form of "alg"?

The thing is that we have a repository manager developed for users being stored in an LDAP-server and when searching for swedish users it would be nice if international users could use a- and o-characters instead of ä-, å- and ö-characters.

Thanks,

Pierre

[KarstenH]

Hi Pierre,

you cannot change this by easy configuration. It might be possible via a Python extension to TREX, but is probably not worth the effort. Especially, as many of these cases should be caught by the TREX fuzzy search algorithm, if you do create a fuzzy index for your user search. Although your exmample may be to short a string for fuzziness to have the desired effect. For those cases (fuzziness does not work) you could use the inclusion of synonyms in your search as described here:

http://service.sap.com/~sapidb/011000358700002581292005E/SemSearch.pdf

Or here in SDN, where it had vanished for a while:

https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/library/kmc/how to enable semantic search or search for synonyms in trex.pdf

You would then have to maintain each name though. ä=a will not do it there, the XTM file used (see link) would have to hold the info that älg=alg=aelg.

Regrads,

Karsten