Prevent ingestion of jobs with 0 start or end time

[jtpalmer]

Description

Changes ingestor SQL query to exclude jobs with a start or end time that is 0.

Motivation and Context

Jobs with zero start time should not be included in the data warehouse.

Tests performed

Shredded and ingested data containing jobs with zero start time. Observed that before the change there were rows in modw.jobfact with start_time_ts zero before the change, but none after.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project as found in the CONTRIBUTING document.
• I have added tests to cover my changes.
• All new and existing tests passed.

[jpwhite4]

Could you also add a line or two to the documentation (or FAQ perhaps?) to explain this new (desired) behaviour? I'm anticipating someone complaining that the job count is off!

[jtpalmer]

Documentation has been added to the FAQ.

[smgallo]

For XSEDE, we assume that the end time is the correct value and if start_time > end_time or start_time <=0 we calculate it based on the end _time and wall duration. Should we do something similar here? We still skip jobs with 0 end time.

[jtpalmer]

@smgallo Open XDMoD will calculate a missing start time if both the end time and wall time are present. This PR documents that in the FAQ. I can move that note somewhere else if that would be more appropriate.

Edit: We could also move that logic out of the shredder and into an ingestor if that would be preferable.

[smgallo]

Lets leave the logic where it is for now. As we move from the old ETL to ETLv2 we can see if it makes sense to move it elsewhere.