Orangutan » Orangutan::Redmine

In my project all time entries were made with Orangutan and chart does shows the corresponding issues (while not all logged hours are shown).
So I believe the problem is not in Orangtan but rather in some specific data in some time entries...

Ok. This somehow depends on chart_time_entries table. Orangutan writes nothing to this table! Currently I’m not sure if it should... It looks like Redmine does it by itself (some Orangutan hours got there others didn’t).

This is the query Redmine uses to get statistics for chart:

SELECT
  SUM(logged_hours) AS logged_hours,
  SUM(entries) AS entries,
  chart_time_entries.issue_id AS group_id,
  issue_id AS grouping,
  issues.estimated_hours AS estimated_hours,
  issues.subject AS subject,
  chart_time_entries.issue_id AS root_id,
  null AS parent_id
FROM chart_time_entries
LEFT JOIN issues ON issues.id = issue_id
WHERE 
  (
    chart_time_entries.month = 0 AND
    chart_time_entries.week = 0 AND
    chart_time_entries.day = 0 AND
    chart_time_entries.project_id IN (23)
  )
GROUP BY chart_time_entries.issue_id, issues.estimated_hours, issues.subject
ORDER BY 1 DESC, 3 ASC

Думаю, Redmine час від часу обновлює вміст таблиці chart_time_entries:

SELECT
  spent_on,
  CONCAT('2010', DAYOFYEAR(spent_on)) AS yday,
  TRUNCATE(SUM(hours), 2) AS hours,
  TRUNCATE((SELECT logged_hours FROM chart_time_entries WHERE issue_id = 246 AND user_id = 34 AND day = yday), 2) AS logged_hours,
  COUNT(*) AS count,
  (SELECT entries FROM chart_time_entries WHERE issue_id = 246 AND user_id = 34 AND day = yday) AS entries
FROM time_entries
WHERE issue_id = 246 AND user_id = 34
GROUP BY yday;

+------------+---------+-------+--------------+-------+---------+
| spent_on   | yday    | hours | logged_hours | count | entries |
+------------+---------+-------+--------------+-------+---------+
| 2010-04-28 | 2010118 |  3.15 |         3.15 |     2 |       2 | 
| 2010-04-29 | 2010119 |  8.00 |         8.00 |     1 |       1 | 
| 2010-05-04 | 2010124 |  3.11 |         3.11 |     1 |       1 | 
| 2010-05-05 | 2010125 |  7.67 |         7.67 |     2 |       2 | 
| 2010-05-06 | 2010126 |  8.00 |         8.00 |     2 |       2 | 
| 2010-05-07 | 2010127 |  7.76 |         7.76 |     3 |       3 | 
| 2010-05-11 | 2010131 |  8.00 |         8.00 |     2 |       2 | 
| 2010-05-12 | 2010132 |  7.91 |         7.92 |     2 |       2 | 
| 2010-05-13 | 2010133 |  7.88 |         7.88 |     3 |       3 | 
| 2010-05-14 | 2010134 |  7.92 |         7.92 |     1 |       1 | 
| 2010-05-17 | 2010137 |  7.92 |         7.92 |     1 |       1 | 
| 2010-05-18 | 2010138 |  5.90 |         5.90 |     2 |       2 | 
| 2010-05-19 | 2010139 |  6.71 |         6.71 |     3 |       3 | 
| 2010-05-20 | 2010140 |  6.42 |         6.42 |     2 |       2 | 
| 2010-05-21 | 2010141 |  7.50 |         7.50 |     2 |       2 | 
| 2010-05-25 | 2010145 |  7.98 |         1.95 |     3 |       1 | 
| 2010-05-26 | 2010146 |  7.65 |         NULL |     3 |    NULL | 
| 2010-05-27 | 2010147 |  7.76 |         NULL |     5 |    NULL | 
| 2010-05-28 | 2010148 |  6.50 |         NULL |     2 |    NULL | 
| 2010-05-31 | 2010151 |  4.61 |         NULL |     2 |    NULL | 
| 2010-06-01 | 2010152 |  7.65 |         NULL |     2 |    NULL | 
| 2010-06-02 | 2010153 |  8.00 |         NULL |     3 |    NULL | 
| 2010-06-03 | 2010154 |  4.26 |         NULL |     1 |    NULL | 
| 2010-06-04 | 2010155 |  7.95 |         NULL |     4 |    NULL | 
| 2010-06-07 | 2010158 |  5.90 |         NULL |     1 |    NULL | 
+------------+---------+-------+--------------+-------+---------+

Цікаво, на яких умовах це робиться... Можливо і варто щось таке робити в Орангутані...

Потрібна допомога, щоб зрозуміти що воно робить:

vendor/plugins/redmine_charts/lib/redmine_charts/time_entry_patch.rb

Треба чи не треба Орангутану “оновлювати” відповідні записи таблиці chart_time_entries?

Таки треба...

A query to detect incorrect/missing entries in chart_time_entries table:

SELECT
  (YEAR(spent_on) * 1000 + DAYOFYEAR(spent_on)) AS day,
  (YEAR(spent_on) * 1000 + WEEKOFYEAR(spent_on)) AS week,
  (YEAR(spent_on) * 1000 + MONTH(spent_on)) AS month,
  ROUND(SUM(hours), 2) AS logged_hours,
  logged_hours,
  COUNT(time_entries.id) AS entries,
  entries,
  time_entries.user_id,
  time_entries.issue_id,
  time_entries.activity_id,
  time_entries.project_id
FROM time_entries
LEFT JOIN chart_time_entries
ON
  day = (YEAR(spent_on) * 1000 + DAYOFYEAR(spent_on)) AND
  chart_time_entries.user_id = time_entries.user_id AND
  chart_time_entries.issue_id = time_entries.issue_id AND  
  chart_time_entries.activity_id = time_entries.activity_id AND
  chart_time_entries.project_id = time_entries.project_id
GROUP BY
  spent_on,
  time_entries.user_id,
  time_entries.issue_id,
  time_entries.activity_id,
  time_entries.project_id