I'm wondering if there's any utility in returning the first term of an un-stored but indexed field. It wouldn't be much use in a field that broke the input stream up, but what about a "string" type? Or something with KeywordTokenizer?
We could save some index storage space, maybe speed up things a bit. Especially in a "string" type, there's no difference between the indexed term and the stored term, so why use both? My original notion was to let users include an indexed-but-not-stored field in the fl list, and leave up to them to use it wisely. When specified, the behavior would be "return the first term encountered in an indexed-but-not-tokenized field with no guarantees". But it also occurs to me that we could do this automagically for "string" types. I haven't thought this through much, for instance what about specifying fl=*? And this changes current programs, users who already (by mistake probably) specify unstored fields in to be returned would start seeing new data. I'm sure there are other gremlins. Not to mention trying to explain it. And, in truth, this really feels like one of those possible-but-not-worth-the-effort ideas. But it was in my head when I woke up this morning and I thought I'd toss it out there. Erick --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
