0

According to mysql full text search (when you index your table with according properly) and lucene, they all use same algorithm for relevancy. TF-IDF with full reverse indexing. However, comparing the speed in text search between lucene and mysql, there is tremendous differences between them in terms of Relevancy and the speed. What are the main reasons for it ?

[1] https://dev.mysql.com/doc/internals/en/full-text-search.html

met.in
  • 173
  • 1
  • 9
  • They are implemented differently. It is algorithm not a set piece of code. There will always be differences in speed etc. – Namphibian May 19 '14 at 03:12
  • but you wouldnt have 200 times speed differences between themi since even the implementations are different. – met.in May 19 '14 at 05:23
  • Yes u could. A bad implementation is a bad implementation no matter what language or tool is used. Also mysql is a database with full text indexes added later. Lucene is a more of full text index server built to index information. – Namphibian May 19 '14 at 05:30
  • We are talking very mature program written with professional programmers and supported by the forbes 500 :) i dont think this is the case. – met.in May 19 '14 at 06:21
  • the fact remains MySQL was never designed to be a full text search engine. This feature was only recently added. It is not matured as Lucene. Lucene is a very different beast as it is a not a database. You cannot compare the two. MySQL only got full text indexes in version 3.23.23 onwards and it has to share the memory and other resources with database operations. Lucene just indexes. – Namphibian May 19 '14 at 06:28
  • I do rather disagree with the idea that [anything Apache does](http://en.wikipedia.org/wiki/List_of_Apache_Software_Foundation_projects), Oracle could do better. – femtoRgon May 19 '14 at 15:40

0 Answers0