Hi all,
I've got a table with a fulltext index. I don't quite understand the difference between the following queries.
This is what one would expect:
hdbsql HDB=> select snippets(canonical) from sentiment_entity where contains(canonical, 'found -Obama', exact) limit 1;
| SNIPPETS(CANONICAL |
| ------------------------------------------------- |
| and spending, Republicans have often <b>found</b> |
This one should return "found" verbatim without Obama:
hdbsql HDB=> select snippets(canonical) from sentiment_entity where contains(canonical, '"found" -Obama', exact) limit 1;
| SNIPPETS(CANONICAL |
| --------------------------------------------------------------- |
| said he <b>found</b> <b>Obama</b>'s inaugural speech "beautiful |
And so should this one:
hdbsql HDB=> select snippets(canonical) from sentiment_entity where contains(canonical, '"found" -Obama', linguistic) limit 1;
| SNIPPETS(CANONICAL |
| ------------------------------ |
| to find an <b>Obama</b> magnet |
Is this behaviour correct and expected?
-- Micha