On 4/21/2014 2:54 PM, Smith, Mike wrote:
I am writing an rpg program connecting to sql server that joins several files together. One of the files has or potentially can have multiple records and I only want to select the record with the most current hold date.


select a.JOB_ID,a.DESCRIPTION, d.receiveddate, max(b.HOLD_DATE) from
workflowmanager.gisowner.jtx_jobs a left outer join workflowmanager.gisowner.jtx_job_holds b
on a.Job_id = b.Job_id left outer join workflowmanager.gisowner.JTX_HOLD_TYPES c
on b.HOLD_TYPE_ID = c.id join workflowmanager.gisowner.wmx_workflow d
on a.job_id = d.job_id
where a.job_Type_ID = 1
and (convert(char(10),a.End_date,120) >= '2014-01-01'
and convert(char(10),a.End_date,120) <= '2014-03-31' )
and convert(char(10),a.end_date,120) > convert(char(10),a.due_date,120)
group by a.job_id, a.description, d.receiveddate, b.hold_date

The file names are slightly different, but its simply because sql server is making me add the database name.
I've also only included a few of the fields I actually need for testing purposes.
I still get multiple records from gisowner.jtx_job_holds but I'm not sure why

JOB_ID DESCRIPTION receiveddate (No column name)
32002 2130537 2013-12-24 12:00:00.0000000 NULL
32003 2130430 2013-12-23 12:00:00.0000000 2014-01-13 14:56:13.0000000
32004 2130433 2013-12-23 12:00:00.0000000 2014-01-13 14:31:30.0000000
32004 2130433 2013-12-23 12:00:00.0000000 2014-01-13 14:51:18.0000000

Drop b.hold_date from your GROUP BY.
To see why, query only JTX_JOB_HOLDS twice:

select job_id, max(hold_date) from jtx_job_holds group by job_id;
select job_id, max(hold_date) from jtx_job_holds group by job_id, hold_date;


This thread ...


Return to Archive home page | Return to MIDRANGE.COM home page