mismatch between DESI and SDSS galaxy redshifts

[moustakas]

Originally reported by @pbiswas82 in desihub/redrock#254:

We have noted mismatches in the redshifts given in the zall-pix-edr-vac.fits catalogue (DESI catalogue) when compared 
to the SDSS spectroscopic catalogue. We cross-matched the GALEX-SDSS-WISE LEGACY CATALOG (GSWLC) with 
sources in zall-pix-edr-vac.fits and noticed that some of the redshifts given in the DESI catalogue are drastically different 
than that in the GSWLC (which uses SDSS redshifts). In the following figure, we can see this.

Following I have listed the DESI target ID, DESI redshift and redshift from the GSWLC for the four galaxies 
marked in the above figure:
1.
Target ID: 39633240163617085
DESI - z: 1.6529383334623118
GSWLC - z: 0.1165
Link to the spectra: https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633240163617085
2.
Target ID: 39633304151917538
DESI - z: 1.671210092601552
GSWLC - z: 0.0711
Link to the spectra: https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633304151917538
3.
Target ID: 39633345683918575
DESI - z: 1.2412686596561762
GSWLC - z: 0.0418
Link to the spectra: https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633345683918575
4.
Target ID: 39633339031750728
DESI - z: 1.657199768022857
GSWLC - z: 0.0107
Link to the spectra: https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633339031750728

If we go to the spectra links given above, we can examine the DESI spectrum. And if we put the redshift given 
by GSWLC, we will see that spectral lines are fitting correctly. Also, if we see these four objects in the Legacy 
Sky viewer (given on the right side in the spectra linked above), we will be able to understand from size and 
brightness that they are nearby galaxies, not galaxies at z> 1.