Query Handling

Metamorph queries in version 6 are handled more seamlessly from a user's perspective. Words are case-insensitve across the entire Unicode 5.1 locale-independent character set, not just the limited or non-existent 8-bit locale supported by the local operating system. UTF-8 is supported directly, and does not need to be translated to an 8-bit character set for searching. Accents, diacritical marks (umlauts etc.) and full-width vs. half-width differences are ignored by default, and ligatures are expanded. All of these features are tunable; see the <apicp textsearchmode> setting (here) for example.

Wildcards with a suffix bind to a single word as expected in version 6. For example, the query "st*ion" matches "stallion" but not "stuff an onion" nor "stationary"; see the SQL settings wildoneword and wildsufmatch for details. A within-operator query (w/N) counts words instead of characters, and the count is the overall length of the match, not a left-and-right radius from an arbitrary term. Thus searching for "clinton bush obama w/10" will find those terms within a single 10-word span. See the <apicp withinmode> setting (here) for details.


Copyright © Thunderstone Software     Last updated: Apr 15 2024
Copyright © 2024 Thunderstone Software LLC. All rights reserved.