Escaping hyphens in tsquery (postgresql)

user_1151392 0
Viewed 1182

I need to create a tsquery based on the string of the following format:

something-smth-somthing-etc-etc

Calling to_tsquery('something-smth-somthing-etc-etc') returns:

'something-smth-somthing-etc-etc' & 'someth' & 'smth' & 'somth' & 'etc' & 'etc'

Clearly, the string undergoes tokenization, stemming etc. But in our case, the column on which we are making the FTS, already contains tsvector which consists of a single lexeme: 'something-smth-somthing-etc-etc'.

The query select * from sometable where searchee @@ to_tsquery('something-smth-somthing-etc-etc') returns no results.

How can I call to_tsquery, so it will not analyze the provided string and create a single lexeme query?

Or am I missing something more major here?

full-text-search postgresql
1 Answers

a suming you have tsvector with mentionned value, it was not processed, but just inserted as tsvecor type:

t=# select to_tsvector('something-smth-somthing-etc-etc'), 'something-smth-somthing-etc-etc'::tsvector;
                                    to_tsvector                                    |             tsvector
-----------------------------------------------------------------------------------+-----------------------------------
 'etc':5,6 'smth':3 'something':2 'something-smth-somthing-etc-etc':1 'somthing':4 | 'something-smth-somthing-etc-etc'
(1 row)

thn indeed you would have false:

t=# select 'something-smth-somthing-etc-etc'::tsvector @@ to_tsquery('something-smth-somthing-etc-etc');
 ?column?
----------
 f

to hack it you can skip processing on tsquery as well:

t=# select 'something-smth-somthing-etc-etc'::tsvector @@ 'something-smth-somthing-etc-etc'::tsquery;
 ?column?
----------
 t
user_53159740
Related