In article , Eric Flesch
writes:

This is part 2 of anomalies I've encountered in cleaning QSO data.
99 44/100 % is fine. This is the story of the remaining 0.56%.

It looks like you have put some detailed work into this.

The task in producing the "million quasars" catalog is to find the
right optical object for each quasar. If the data positions a
brightish quasar but I find no optical object there, then I must
search for it. Any feedback welcome.

What fraction of quasars are not detectable in the optical at all with
current observations?

Have you contacted the authors of the erroneous data?

Especially if you have an obvious explanation for the mistake it might
be worth writing this up as a short paper. One of the disadvantages of
online data is that it is not obvious when something changes and one
cannot continuously check it for updates. Worse, the authors might
never update their data even when it has been shown to be wrong. Such a
paper might be an important heads-up for folks working with such data.

1) HE 0435-1223 (this old notation describes a B1950 tile of sky):
a Cyril Hazard quasar -- he rarely published positions for these.
Veron positions this at J043814.8-122314, NED has J043814.8-121715.
NED is right -- Veron failed to convert the declination from B1950 to
J2000.

HE is Hamburg-ESO, right? And the other objects have similar names
denoting who discovered them etc. Have you checked the original
literature, rather than compiled catalogues?