wiki:db/queries/021
Last modified 7 years ago Last modified on 02/01/2013 06:20:30 AM

Return moving objects whose motion uncertainty is smaller than measured motion

SELECT objectId
FROM   Object
WHERE  SQUARE(muRA_PS)+SQUARE(muDecl_PS) >
       SQRT (SQUARE(muRA_PS*muRA_PS_Sigma)+SQUARE(muDecl_PS*muDecl_PS_Sigma))

Original query from SDSS:

SELECT  objectId,
        rowC,colC,rowV,colV,rowVErr,colVErr,
        flags,
        psfMag_u,psfMag_g,psfMag_r,psfMag_i,psfMag_z,
        psfMagErr_u,psfMagErr_g,psfMagErr_r,psfMagErr_i,psfMagErr_z
FROM   MovingObject
WHERE  rowvErr > 0 and colvErr> 0 
  AND ((rowV * rowV) / (rowVErr * rowVErr) + (colV * colV) / (colVErr * colVErr) > 4)

This is Moving Object related query. It needs work.

Comment from SDSS: this query takes a long time to run without the "TOP 1000".

This query originally came from SDSS (SX-2)

LJ-

I don't know what this query means in LSST world. In SDSS I can imagine it's relevant because you only have one observation per object, with a very slight time difference from start of observation (whatever filter is at the top of the camera) to the end (whatever filter is at the bottom of the camera).

Translating this equivalently to LSST, I think that means "give me all tracklets which have a velocity greater than the uncertainty in the velocity" ... which would be almost every tracklet. (the uncertainty in the tracklet would come from the uncertainty in the astrometric positions?). This would mean that most of what was returned by this query would be diasource/diasource linkages, rather than true potential moving objects.

We should find out if anyone will actually ever want to do this .. given how fast the number of tracklets will build up with time, I had imagined we might not store them for longer than a given time period, which might make supporting this query difficult.