Question

Given a big table of data about when people begin and complete tasks, e.g:

Person  |  Task | Date started | Date ended
---------------------------------------------
A        Cleaning  20-FEB-2012  22-FEB-2012
N        Dishes    20-FEB-2012  24-FEB-2012
Z        Cleaning  21-FEB-2012  23-FEB-2012

and a score table which assigns scores of 2,3,4 for each task based on how long it takes them to do it, e.g.:

|  Task | Days taken | Score
---------------------------
Cleaning      2        2
Cleaning      1.5      3
Cleaning      1        4
Dishes        3        2
Dishes        2.5      3
Dishes        2        4

how might I produce a query which gives the overall score for each person for each task, e.g.:

Person  |  Task | Overall Score
---------------------------------------------
A        Cleaning  3.1
A        Dishes    2.7
N        Cleaning  3.4

The solution’s been subtly eluding me, some assistance would be appreciated! I’m using SQLite at present.

Answer 1

Your definitions are a bit vague. However, the following should help you:

select t.person, t.task, sum(s.score)
from tasks t left outer join
     score s
     on t.task = s.task and
        s.daysTaken = t.julianday(dateended) - t.julianday(datestarted)
group by t.person, t.task

Handling ranges a bit more difficult. You need to get the two ends of the interval, and then do the join:

select t.person, t.task, sum(s.score)
from tasks t left outer join
     (select s.*,
             (select min(days_taken) from score s2 where s2.person = s.person and s2.task = s.task and s2.days_taken > s.days_taken
             ) as nextDays_Taken
      from score s
     ) s
     on t.task = s.task and
        t.julianday(dateended) - t.julianday(datestarted) >= s.days_taken and
        t.julianday(dateended) - t.julianday(datestarted) < nextDays_Taken
group by t.person, t.task