Problems like this depend heavily on example what the fieldtype and "index" analyzer is for the field you are querying on. it's important to keep in mind that wildcard and fuzzy queries are not "analyzed" so things like lowercasing and stemming have to be taken into account -- typically it's useful to use copyField to have a special version of your field with simplified analysis for doing wildcard searches on.
as for your specific problem: given the limited information you've provided, no guesses immediately jump out at me as to what you should od to get things working the way you want ... it depends on your schema, and what the orriginal text was in those 3 documents you want to match. : For example, following are the search queries and the corresponding results : tomcat* -> 3 results : tomca* -> 0 results : tom*at -> 0 results : tom~at -> 0 results -Hoss
