i'm trying build query measures "trending" hash tags tweets. have table tweets , column message (varchar)
basically want group messages hash tag breaks down if there more 1 hash tag in message field, heres have far
select count(message) count, `message` `tweets` `message` rlike '^#[[:alnum:]]' group `tweets`.`message` order `count` desc
which output following rows
"4","#sushi" "3","#coffee" "3","#wine" "1","#coffee #park" "1","#drinking" "1","#steak" "1","#pizza" "1","#pasta" "1","#food"
the fourth row problem want both hash tags totaled on own, not sure if can achieve mysql.
your column message
want group hashtags. database design pov, table isn't in 1nf because message
holds repeating group of hash tags. won't able use power of sql on table until that's corrected.
what need table hashtags
has 1 row every tag in every message. once have that, query practically writes itself.
Comments
Post a Comment