I'm working on a production recording database for our plant.
One of the metrics I need to measure is downtime. Downtime
can fall into 3 different buckets:
1) Unplanned - Related to the production run
2) Unplanned - Not related to the production run
3) Planned - Not related to the production run

Unplanned could be a mechanical breakdown, while planned would be
something like a safety meeting.

Production runs are supplied a unique run number, so some of my downtime will be assigned to a run number, and some of it won't.
I'm thinking of a table like this:

DownTime
========
Date
Shift
ReasonCode
RunNumber (allow nulls)
Minutes

Where date, shift, and reason code would be composite key.
For some reason I feel like I'm missing something, though. Does anyone
see a better way to do it?

Thanks
Mark

Date/Shift/ReasonCode should not be a primary key, as I would assume downtime could possibly occur twice on a shift for the same reason. I think you just need a timestamp in there instead to complete the key.

That's true. I wasn't thinking about that. I think I'll go with a start and stop time instead of total minutes.
Thanks