Simplier count(*)

[rodo]

In
https://github.com/jbranchaud/til/blob/master/postgres/count-records-by-type.md

I prefer select type, count(*) from table group by 1 as is if the elements counted are something like select substring(unnaccent(upper('firstname',1,2))) .... there are less character to type at the end of the command :)

[jbranchaud]

@rodo good tip! I also try to use the indexes when ordering.

[rodo]

It's just an alias, here if you have an index on type it will be used.