quickquasars sets every EXPID=0

[ngbusca]

quickquasars sets all the EXPID to zero. This is a problem for picca since pairs at rp == 0 are rejected. While we improve this on the picca side (not before DR14 papers are out) it would be good to fix this on quickquasars too.

I've included a "hack" in picca to avoid this issue (currently in branch picca_nersc_submit) by setting the expid to the nightid (the latter is different for each line of sight).

[andreufont]

Is EXPID the equivalent of FIBERID, or the equivalent of THINGID? I'm guessing FIBERID?

[ngbusca]

There's no exact equivalent when using coadded data. Picca was using it as the PLATEID, meaning that pairs with equal EXPID and both FIBERID < 500 (trivially satisfied for realistic desidata but not for quickquasars where FIBERID can be very large) and rp==0 are rejected.

[ngbusca]

By the way, the NIGHT is also the same for all quickquasars spectra-16 files, so I've set the "effective" plate in picca as the FIBERID instead of the NIGHT.