How Fast is Oracle at Indexing Text and Can I Just Enable Boolean Searches?
We are using mySQL to do partial indexing of 9 million Web pages a day. We are running on a 4-processor Sparc 420 and unable to do full text indexing. Can Oracle8i or Oracle9i do this? We are not interested in transactional integrity, applying any special filters to the text pages, or in doing any other searching other than straight boolean word searches (no scoring, no stemming, no fuzzy searches, no proximity searches, and so on). I have are the following questions: Will Oracle be any faster at indexing text than mySQL? If so, is there a way to disable all the features of text indexing except for boolean word searches? Answer Yes. Oracle Text (interMedia Text) can create a full-text index on 9 million web pages – and pretty quickly. In a benchmark on a large Sun box, we indexed 100Gig of web pages (about 15 million) in 7 hours. We can also do partial indexing via regular DML or via partitioning. You can do “indexing light” to some extent – you can disable theme indexing, you do not n
