Re: MkIV database - duplicate data

[Michael Sallman <msallman@pro-ns.net>]

Tom,

> Well, I try to keep all this straight.  I am still evolving a
> bookkeeping scheme.  So this is possible.  But I did send all the .cal
> files.  So a duplicate name check *could* catch this.  It could also
> fail.  One has to search for a duplicate julian day.  Depending on the
> cloud problems, a frame could be skipped at the beginning of the run.
> If I processed a day twice, and with a different cloud rejection scheme,
> then it is possible to have two .cal files from the same day with
> slightly different names.  The check is not to allow two .cal files from
> the same telescope and the same day.

I'm sure there are other issues that we will run into in the future. You 
  just hope you can anticipate as many as possible at the start :-)
I never thought to check for duplicates before loading the data. I will 
be able to do this going forward.
Right now I'm querying the database for any objects that have multiple 
observations with the same JD. We'll see what is found.

> For some time I have been processing a day at a time.  I see no reason
> that I will not continue this.  Earlier I was processing a disk pair at
> a time.  This would give more than one .cal file a day.
>
> I propose that all the sites generate one .cal file per evening per
> system of processed stars.  This should then allow pretty simple
> bookkeeping to keep everything straight.

I agree. This seems the best way to go.

Mike