Question

I have an attendance table that I’m trying to query in order to create a dataset that I can graph with. My table stores attendance information like so:

date       | student_id | present | session |
-----------|------------|---------|---------|
2012-09-01 | 1          | 1       | AM      |
2012-09-01 | 1          | 1       | PM      |
2012-09-01 | 2          | 0       | AM      |
2012-09-01 | 2          | 1       | PM      |
2012-09-02 | 1          | 0       | AM      |
2012-09-02 | 1          | 0       | PM      |
2012-09-02 | 2          | 0       | AM      |
2012-09-02 | 2          | 0       | PM      |

I’m trying to end up with a dataset that has each unique date, along with the number of present marks (divided by two, eventually, for a kinda accurate representation of attendance for each day).

2012-09-01 | 3 (1.5)
2012-09-02 | 0 (0)

This is so far the closest I’ve got

AttendanceMark.group(:date).count(:conditions => "present = 1")

but this resulting hash doesn’t include the dates where there are no present marks at all. How would I go about achieving this?

Answer 1

If you are using 0 and 1 for present and absent respectively, you can construct a query like this

# For MySQL
AttendanceMark.group(:date).sum('CASE present WHEN TRUE THEN 1 ELSE 0 END')

# For SQLite
AttendanceMark.group(:date).sum('CASE WHEN present = "t" THEN 1 ELSE 0 END')

This shall return

{ "2012-09-01" => 3, "2012-09-02" => 0 }

You can later divide all the sums by two.