Question

I’m trying to perform calculations, apply conditions for the calculated values and count results all in one query. The problem is that I don’t know how to refer to the calculated field from the conditions.

I have Issues.created_on field and I want to find statistics about how many issues were created each week during a range of weeks provided by user.
This how my query looks like in MySql:

mysql>  SELECT status_id as group_id,
YEAR(created_on)*1000+WEEK(created_on,1) as range_value, count(*) as
noOfEntries FROM `issues` WHERE (`issues`.`project_id` IN (1))  GROUP
BY issues.status_id, range_value;

+----------+-------------+-------------+
| group_id | range_value | noOfEntries |
+----------+-------------+-------------+
|        1 |     2012031 |           2 |
|        5 |     2012015 |           1 |
+----------+-------------+-------------+

Here is my code (simplified):

# Range value is dynamic: "YYYYXXX", where YYYY is year and XXX could be week, month or day of year
range_value = "YEAR(created_on)*1000+WEEK(created_on,1)"    
select = "#{range_value} as range_value, count(*) as noOfEntries"
grouping = "range_value"
# other conditions are provided by user, so we just add one here
conditions["range_value"] = range[:min]..range[:max]

rows = Issue.all(:select => select, :conditions => conditions, :readonly => true, :group => grouping)

But I get this error:

ActiveRecord::StatementInvalid (Mysql::Error: Unknown column 'issues.range_value' in 'where clause': SELECT YEAR(created_on)*1000+WEEK(created_on,1) as range_value, 'weeks' as range_type, count(*) as logged_hours, 1 as entries, 'priority_id' as grouping, issues.priority_id as group_id FROM `issues` WHERE (`issues`.`range_value` BETWEEN '2012012' AND '2012031' AND `issues`.`project_id` IN (1))  GROUP BY range_value, issues.priority_id ORDER BY 1 asc, 6 asc):

The code I’m trying to modify is not mine, so I cannot add any new fields or methods to the Issues object. So solution like this will not work for me, I guess.

Answer 1

I have found an example of code which actually helped:

ActiveRecord doesn’t have a parameter for including a HAVING clause in
your database queries and I sometimes find myself needing it. Luckily,
you can sneak it in on the end of your :group parameter without
needing to resort to a find_by_sql() call. (I don’t recall ever using
HAVING without GROUP BY, though there probably are some cases where it
would make sense to do so.)

As an example, here’s a query from one of my Rails applications that
finds all duplicate email addresses in the database:

duplicates = User.find( :all,
  :select     => "email, COUNT(email) AS duplicate_count",
  :conditions => "email IS NOT NULL AND email != ''",
  :group      => "email HAVING duplicate_count > 1"
)

So after a slight modification my code works:

# Range value is dynamic: "YYYYXXX", where YYYY is year and XXX could be
# week, month or day of year
range_value = "YEAR(created_on)*1000+WEEK(created_on,1)"    
select = "#{range_value} as range_value, count(*) as noOfEntries"
# CHANGED THIS LINE from: grouping = "range_value"
grouping = " HAVING range_value BETWEEN #{range[:min]} AND #{range[:max]}"

# other conditions are provided by user, so we just add one here
# REMOVED THIS LINE
# conditions["range_value"] = range[:min]..range[:max]

rows = Issue.all(:select => select, :conditions => conditions, :readonly => true, :group => grouping)