ippToPsps SA8

This iteration of SAS involves many changes in both the dvo schema and the psps schema. Some things are not implemented yet, and some have problems. This wiki lists the problems (and possible solutions).

dvo

Here are the ingredients for the dvo:

• cam stage = 'SAS_v8'

• skycal stage = 'SAS_v8' - note that this is the first time we have ingested the skycal stage. This replaces the staticsky stage

• dvo label = 'SAS.20120801'
• the dvo is type PS1_v4, with many new columns in objects
• there are 620 camera stage and 2849 skycal stage ingested.
• there are slightly fewer skycals ingested than the number of staticsky ingested on SA7 - this is due to crankiness on staticsky's part.

ipptopsps

there are many changes to ipptopsps between sa7 and sa8, all in the object tables:

• many columns have been renamed and changed, see ​http://svn.pan-starrs.ifa.hawaii.edu/trac/ipp/wiki/ippToPsps_Object the new changed columns:
  • xMeanPSFMag/Err - relating to the mean psf mag of the camera stage detections (previously xMeanPSF)
  • xMeanKronMag/Err - mean kron mag/err of the camera stage detections.
  • x20pct/x80pct - previously was xMin/xMax - renamed to be more precise
  • xStackPSFFlux/Err - the stack psf flux/err
  • xStackKronMag/Err - the stack kron mag err
  • xFlags - info flags for the different filters
• due to restrictions, we just renamed currently unused or easy to calculate columns. Most of these involve color, or ill defined columns like magbest.

known issues

• sgsep, xflags is not populated. There's no particular reason why this was missed other than there were too many columns for Heather to keep track of

• x20pct/x80pct default to 32.767 - this is a bug in ipptopsps - it didn't know this was NAN, it will replace those with -999 in the next iteration

• there is garbage xMeanKronMag and xMeanKronMagErr when nx = 0. xMeanPSFMag and xMeanPSFMagErr properly default to -999
  • this comes directly from the dvo, so likely should be fixed there

• xStackKronMag is off by an offset (needs to add 17.8 to get the correct value). This is a typo in ipptopsps, future SAS won't have this

• top 10 * from objects lists 10 objects with -999 for all fields except ra and dec (no mags at all, no flags at all.. why are we publishing those? I still need to investigate, but I suspect dvo).

• how to grab good stuff? there is still discussion on this. some ideas (note that we intend to fix future SAS so that these will not be needed):
  • select top 10 * from Object where ng > 0 and nr > 0 and ni > 0 and nz > 0 and ny > 0
    • this reduces the number of objects from 16M to 3M.
  • select top 10 * from Object where nr > 0 and ni > 0