Rails - Active Record: add extra select column to find(:all)

Question

I want to add new column sum(time_entries.hours) to sql select I'm querying entries like this:

issues = Issue.visible.where(options[:conditions]).all(
    :include => ([:status, :project, :time_entries] + (options[:include] || [])).uniq,
    :conditions => statement,
    :order => order_option,
    :joins => query_joins(order_option.join(',')),
    :limit => options[:limit],
    :offset => options[:offset],
    :group => "#{Issue.table_name}.id"
)

It generates this select:

SELECT
 "issues"."id"                         AS t0_r0,
 ...
 "time_entries"."hours"                AS t3_r4,
 ...
 "versions"."sharing"                  AS t8_r9
FROM "issues"
 LEFT OUTER JOIN "projects" ON "projects"."id" = "issues"."project_id"
 LEFT OUTER JOIN "issue_statuses" ON "issue_statuses"."id" = "issues"."status_id"
 LEFT OUTER JOIN "time_entries" ON "time_entries"."issue_id" = "issues"."id"
 LEFT OUTER JOIN "users" ON "users"."id" = "issues"."assigned_to_id"
 LEFT OUTER JOIN "trackers" ON "trackers"."id" = "issues"."tracker_id"
 LEFT OUTER JOIN "enumerations"
   ON "enumerations"."id" = "issues"."priority_id" AND "enumerations"."type" IN ('IssuePriority')
 LEFT OUTER JOIN "issue_categories" ON "issue_categories"."id" = "issues"."category_id"
 LEFT OUTER JOIN "versions" ON "versions"."id" = "issues"."fixed_version_id"
WHERE "issues"."id" IN (1) AND (projects.status <> 9 AND projects.id IN (SELECT
                                                                       em.project_id
                                                                     FROM enabled_modules em
                                                                     WHERE em.name =     'issue_tracking')) AND
  ((issues.status_id IN (SELECT
                           id
                         FROM issue_statuses
                         WHERE is_closed = 'f')))
GROUP BY issues.id
ORDER BY issues.id
 DESC

I tried insert :select=>"*, sum(time_entries.hours)" into hash parameter to all method, but there is no effect.

How can I add new column to this select? Also I need preserve all current columns, since they are used in filters. Is there any way without specifying all columns by hands?

UPDATE: Ruby 1.9.3 Rails 3.2.13

Answer 1

The table names look like from Redmine. I actually ran into a similar problem recently. What I found was that the selected colums stem from the visible scope. I could not pin down why exactly but working without it, the 'select' option worked as intended.

Using

Issue.select("#{Issue.table_name}.*, sum(#{TimeEntries.table_name}.hours AS total_hours").where("#{your_conditions}").group(:id).all

should give you the data you want. The additional column can be accessed directly as an Issue attribute:

Issue.total_hours

Edit:

If it's not possible to select all Issue fields (for creating the full Issue objects) and the sum at the same time, you need to use a second statement after getting the issues you want.

issues = Issue.visible.where("#{issue_conditions}").all
issue_ids = issues.map(&:id)
# if you do not need the issue objects, use .pluck(:id) instead of .all

spent_hours = TimeEntries.select('issue_id, sum(hours)')
                         .where(:issue_id => issue_ids)
                         .where("#{other_time_conditions}")
                         .group(:issue_id).all

---

Alternatively you can use the included convenience functions in Redmines Issue class. Use this to make the Software do the work for you:

i = Issue.includes(:time_entries).find(1)
i.total_spent_hours

The include makes sure that the calculation doesn't hit the database a second time because all related associations are already loaded.

Answer 2

Bit old but something like this should work. (In Rails 5 ish)

issues = Issue. ...

issues.select(
  issues.arel.projections,
  TimeEntry.arel_table[:hours].sum
)