> I am wondering if there is some way to maintain a stopword > list with widcards: > > ignoring anything that starts with "foo": > foo*
A custom TokenFilterFactory derived from StopFilterFactory can remove a token if it matches a java.util.regex.Pattern. List of patterns can be loaded from a file in a similar fashion to stopwords.txt > i am doing some funky hackery inside DIH via javascript to > make my autosuggest work. i basically split phrases and > store them together with the full phrase: > > the phrase: > "Foo Bar" > > becomes: > > Foo Bar > foo bar > {foo}Foo_Bar > {bar}Foo_Bar What is the benefit of storing {foo}Foo_Bar and {bar}Foo_Bar? Then how are you querying this to auto-suggest?