I fear I have an involved challenge concerning FTS.
Assume I have the following text in a column:
Graz,06.Bez.:Blah
This parses as: SELECT alias, description, token FROM ts_debug('german', 'Graz,06.Bez.:Blah'); alias | description | token -----------+-----------------+-------- asciiword | Word, all ASCII | Graz blank | Space symbols | , host | Host | 06.Bez blank | Space symbols | .: asciiword | Word, all ASCII | Blah
Bez. ist the abbreviation for "Bezirk" (german for ~district). 06.Bez means "6th district"
My first problem might be that the parser identifies "06.Bez." as a host lexeme, but ...
I already defined a synonym dictionary to enable searching for "Bezirk", when there is only "Bez." in the database:
file: bevaddress_host.syn: 01.bez bezirk 06.bez bezirk <snip some more rows>
CREATE TEXT SEARCH DICTIONARY bevaddress_host_syn ( TEMPLATE = synonym, SYNONYMS = bevaddress_host ); ALTER TEXT SEARCH CONFIGURATION german ALTER MAPPING FOR host WITH bevaddress_host_syn, simple;
I wonder how I can achieve to be able to search for "Erster Bezirk" ("First district") to match eg. "01.Bez."
As of the time of writing this I haven't seen any replies to your post so you may not be aware that an answer was provided to your specific question in a blog. http://obartunov.livejournal.com/185579.html