Postgres' full text search
to_tsvector() function returns a sorted array of lexemes. I need exactly that, just with lexemes in original sort order. Is there any such Postgres function, or do I need to do it myself? Is so, how?
I need it for Jaro-Winkler string similarity function, for which order of strings/words is important. I need
tsvector only because of language support for "lexemization". Actually, I just need to convert text, e.g.:
select array_to_string(tsvector_to_array(to_tsvector('english', 'I just want to lexemize this string and keep the original order of lexemes...')), ' ');
This is OK, except lexemes are ordered alphabetically instead of their original relative order in original text.