When in doubt, WhitespaceAnalyzer is the most predictable. Note that it doesn't lower-case the tokens though. Depending upon your requirements, you can always pre-process your query and indexing streams and do your own lowercasing and/or character stripping.
You can always create your own analyzer with the building blocks provided via Filters and Tokenizers.... Erick. On 7/18/07, WATHELET Thomas <[EMAIL PROTECTED]> wrote:
Witch analyser I have to use to find text like this '<commision>'?
