Releases before DP0.3 SSSource::heliocentric[x,y,z] columns seems nonsensical in June 2021 release

[mjuric]

@jrob93 reports via Slack:

Hi all, I think I might have found a problem with heliocentricz and heliocentricvz in the SSSource table. I've been trying to propagate the orbit of an object so that I can resample things like ra and dec in between the sparse observations in the database but the z-coordinates don't seem to make much sense. E.g. z having values ~1AU when the orbit is only a couple degrees inclination. I've put a notebook at ~/shared/jrob93/check_cartesian_coords that should demonstrate the problem.

This is most likely a bug in the coordinate converter code; should be fixed before the 2022 refresh of the catalog. (thanks @jrob93 for reporting it!).

[jrob93]

Hey @mjuric I think this problem has affected more than just the stored z components. I couldn't get phase angles I have been calculating to match up with the stored database values. However, I recover the database phase when I calculate it directly from the topocentric and heliocentric (which has incorrect z components) vectors stored in the database.

So it looks like the problem with the z components is leading to incorrect phase angles as well.