I'm going to establish a large database (ISAM) of simple structure, say the table with only 3 columns:
- VARCHAR(80) - VARCHAR(40) - DATE The number of rows can be quite large, about hundred million or so. The first column, containing actual information, will contain unique values, unlike the two others - but the two others shall be used for data selection (and I'll index them). Now to the point: considering, that the second column shall contain about 100-200 different "labels" - so in average many of such labels can be repeated one million times (or even more) - will it speed-up the selection done with something like "...WHERE label='xyz' AND date='2013-02-25' " (and maybe the insertion as well?), if instead of textual labels I'll use SMALLINT as simple integer "code" for each different "label" (and in my program a translation table)? If so - will I have further performance gain, if instead of human-readable DATE for the third column I'll use "Unix time" put into INTEGER type column, not DATE-type? And the final question: even, if so - is it worthy? I mean: will the supposed performance gain be significant (e.g. 2-3 times faster selection) - and not, say, just 5% faster (only possible to detect by using benchmarking tools)? Thanks in advance for your opinions. -- Zbig -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/mysql