Subject | Re: Google-like scoring in databases |
---|---|
Author | Roman Rokytskyy |
Post date | 2003-07-01T06:08:35Z |
Jim,
Last weekend I successfully integrated Lucene with Firebird via
JayBird (indices are stored in blobs), performance is reasonably good;
indexing is ~20% slower than using file system, searches in many cases
are faster (probably due to Firebird's cache), approx 20-80ms per
search in database with 4,500 documents of JDK 1.3.1 API on my 650MHz
laptop.
Lucene has various text filters that one can use (stop words,
stemmers, etc. and they are BSD licensed, you might want to consider
using them in Netfrastructure - they are pretty generic).
Best regards,
Roman Rokytskyy
Last weekend I successfully integrated Lucene with Firebird via
JayBird (indices are stored in blobs), performance is reasonably good;
indexing is ~20% slower than using file system, searches in many cases
are faster (probably due to Firebird's cache), approx 20-80ms per
search in database with 4,500 documents of JDK 1.3.1 API on my 650MHz
laptop.
Lucene has various text filters that one can use (stop words,
stemmers, etc. and they are BSD licensed, you might want to consider
using them in Netfrastructure - they are pretty generic).
Best regards,
Roman Rokytskyy