Fuzzy search not behaving as expected in all cases? about lokidb HOT 3 CLOSED

Fuzzy uses leventhstein distance (max. distance is 2). All returned results look good to me.

I recommend the use of multiple FTS queries and weight them differently. So try to combine fuzzy and wildcards.

from lokidb.

hi @Viatorus , thx for the quick answer, I was thinking I'm most likely doing something wrong, it's interesting I did not know you can combine the fts queries, I didn't really see a possibility for that in the types, can you give a quick syntax example of how to use multiple queries at once?

from lokidb.

Checkout the documentation of elastic search and the test_cases. The query syntax is very similar.

Untested:

{
$fts: {
  query {
      type: "bool",
      should: [
        {
          type: "wildcard",
          field: "name",
          value: "*" + search.value,
          boost: 2
        },
        {
          type: "fuzzy",
          field: "name",
          value: search.value,
          extended: true,
          prefix_length: 0
        }
      ],
      minimum_should_match: 1
    },
  }
}
}

from lokidb.