Or, very similar, wrap the 'real' analyzer A with your analyzer that delegates to A but also keeps the returned tokens, possibly by using a CachingTokenFilter.
On Jan 7, 2008 7:11 AM, Daniel Noll <[EMAIL PROTECTED]> wrote: > On Monday 07 January 2008 11:35:59 chris.b wrote: > > is it possible to add a document to an index and, while doing so, get > the > > terms in that document? If so, how would one do this? :x > > My first thought would be: when adding fields to the document, use the > Field > constructors which accept a TokenStream and pass in a CachingTokenStream > wrapped around the real token stream. > > Daniel >
