Entries this day: today-worked-on-ab-db

today worked on ab db

##22:32 Monday 16 February 2015 JST##

AB DB primarily consists of two big tables. Table U has 77k records, and 160 columns. Table B has maybe 40k records and maybe 70 columns. There's a 1-to-many relationship from B to U. Doing joins on them takes about 3 or 4 seconds.

That ain't good, considering the joins are done all the frickin time.

I guess that's why when I got the original data, a lot of what's now in the B table was in the U table. I was like, "normalization to the rescue!" but now it's sloooowwwww.

About 20 columns in table U are varchar 200, so that's 2k UTF-8 characters, so 4k bytes per record (if they're all filled, which they're not, but it's still overhead that I wonder if I can split into other tables).