Count similar entries in SQL

Question

I have a table where a column for grade of students.
We have A+ and A and A- and B+ and B and B- .
Now we want to treat A+ and A and A- all equal to A . So as B+ and B and B- all equal to B .

How to write the expression to aggregate this? The question is count how many A and B? A+ and A- also mean A. ?

Answer 1

One way of many:

SELECT left(grade, 1) as base_grade, count(*) AS ct
FROM   tbl
GROUP  BY 1
ORDER  BY 1

left() requires Postgres 9.1+. Else use substring(grade, 1, 1) .