The problem I am struggling with is I have a database of batch jobs,
very simple.

JOBNAME,START TIME, FINISH TIME

JOB1STOP;1/1/7 9:00:00;1/1/7 9:00:00
....
JOB1STOP;1/1/7 9:01:00;1/1/7 9:07:00
....
JOB1START;1/1/7 12:34:01;1/1/7 12:54:03
....
JOB1START;1/1/7 12:37:01;1/1/7 12:59:03

The reality of this is The JOB1STOP begins a database down for
maintenance and JOB1START begins it back up.

The second JOB1STOP doesn't matter (it's already down) nor does the
second JOB1START (it's already up) so I am guessing that a query in
psuedo-code would look like (assuming we want just jobs from 1/1/7:

SELECT (jobname,start time, finish time) from JOBTABLE where STARTTIME
like "1/1/7" AND JOBNAME like "START" or "STOP"

But then I have to figure out how to clip only the first STOP and the
FIRST START ignoring the duplicates. (distinct perhaps?)

Depending on the server the logs are from means I would invert the
criteria needing only the LAST of a given record. (order by prior to
distinct?)

I am not very good with complex SQL queries, but my best guess to this
point is (using $vars)

SELECT (jobname,start_time,end_time) from $TABLE
WHERE start_time >= "1/1/7" AND
jobname LIKE ("START" OR "STOP")
ORDER BY start_time DISTINCT

Am I close? and How could I ignore the query if one of the two are not
present (ignore STARTS or STOPS on a day that the other is missing?)

Id..

Idgarad wrote:
You need what I call the "Strawberry Query"
See the pattern for it in my response on this thread:
http://tinyurl.com/yex7s2