MySQL连接并使用零对查询进行计数

Question

I've been staring at so much SQL code my brain is practically mush.

I'll make it quick. Below are my table structures, with irrelevant columns omitted:

attendance_history:              member_info:
_____________________            ________________________
|   Date   |   ID   |            |   ID   | Design Team |
|----------|--------|            |--------|-------------|
|   1/27   |    1   |            |    1   |     DT1     |
|   1/28   |    1   |            |    2   |     DT2     |
|   1/29   |    2   |            |    3   |     DT2     |
|   1/29   |    3   |            |--------|-------------|
|----------|--------|

My ultimate goal is to get an attendance average for every design team. But I believe I have all of the logic worked out except one part: I need to get the number of attendees for EACH design team for EACH date.

Here is an example of what I need:

__________________________________
|   Date   | Design Team | Count |
|----------|-------------|-------|
|   1/27   |     DT1     |   1   |
|   1/28   |     DT1     |   1   |
|   1/29   |     DT1     |   0   |
|   1/27   |     DT2     |   0   |
|   1/28   |     DT2     |   0   |
|   1/29   |     DT2     |   2   |
|----------|-------------|-------|

Any clues as to how I can achieve this? Currently I have a ridiculously long script that only gets the counts for the design teams if someone was present. I need somehow get a 0 for a count in a row if there was nobody present for that day from that team.

Thanks a ton, as always.

Answer 1

To get the zeros requires a bit of trickery. You need to generate all the rows first (using a cross join ) and then left join in the attendance information:

select d.date, dt.design_team, coalesce(ddt.cnt, 0) as cnt
from (select distinct design_team from member_info) dt cross join
     (select distinct date from attendance_history) d left outer join
     (select ah.date, mi.design_team, count(*) as cnt
      from attendance_history ah join
           member_info mi
           on ah.id = mi.id
      group by ah.date, mi.design_team
     ) ddt
     on ddt.design_team = dt.design_team and ddt.date = d.date;

Edit : Fixed subquery to join on ah.id = mi.id instead of ah.id = ah.id.