I have a table (job_logs
) 开发者_运维百科with the following records:
id
, job_id
, user_id
, status
, created_at
, job_type
.
Each time a job starts to run a record is written in the job_log
table with status='started'
. When a job finish running another record is added to the table with status='completed'
.
Both records has the same user_id
, job_type
and job_id (which is determined by the process running the job - unique to these 2 records).
I want a query that will return all these records pairs in the table (ordered by id desc) but the tricky part is that I want to add to the record with the 'completed' status the time it took the job to run (completed.created_at - started.created_at
).
How can I do that?
SELECT j1.job_id AS job_id, (j2.created_at - j1.created_at) AS time_run
FROM job_logs j1 INNER JOIN job_logs j2 ON (j1.job_id = j2.job_id)
WHERE j1.status = 'started' AND j2.status = 'completed'
精彩评论