[sqlite] glob and like not using index in newest sqlite

2010-11-26 Thread Spiros Ioannou
Hello, I recently updated from 3.4.2 to 3.7.3 and the glob and like operators are now not using the indexes. My table "dict" has a column named "word" without defined type, containing utf-8 words. The index is: CREATE INDEX wordidx on dicts (word); explain query plan SELECT * from dicts where word

Re: [sqlite] LIKE operator extremely slow? more than 10 times slower than egrep?

2007-11-30 Thread Spiros Ioannou
Thank you all for your input. -Spiros -- Image Video & Multimedia Systems Lab. Department of Electrical & Computer Eng. National Technical University of Athens http://www.image.ece.ntua.gr [EMAIL PROTECTED] wrote: Spiros Ioannou <[EMAIL PROTECTED]> wrote: John Stanton wrote

Re: [sqlite] LIKE operator extremely slow? more than 10 times slower than egrep?

2007-11-28 Thread Spiros Ioannou
John Stanton wrote: Sqlite does cater for text searching. Look at FTS. This is not text searching. No stemming, etc etc is required. Column has exactly 1 word, and the 'LIKE' substring matching is performed at the words' first characters (not end-characters). Thanks, -Spiros --

Re: [sqlite] LIKE operator extremely slow? more than 10 times slower than egrep?

2007-11-28 Thread Spiros Ioannou
Firstly, thank you all for your answers, egrep gets to work with a flat text file, which it can easily read sequentially and get optimal performance from the OS's file buffering/cache management. It only needs to read a piece of the file and scan for patterns, repeating until done. The only st

[sqlite] LIKE operator extremely slow? more than 10 times slower than egrep?

2007-11-27 Thread Spiros Ioannou
erformance at all in this case? Is it because non-latin chars are used? Please clarify, I'm really confused about this. Thank you. Spiros Ioannou p.s. if answers to (1) and (2) don't help, I'm going to write a php-egrep plugin to avoid popen-ing :-) -- Image Video & Mult