The issue arises when using xmatch on a small sample of stars. In the example below 7 out of 10 stars queried with xmatch are identified initially. After dropping the previously unidentified stars from the sample, even fewer stars are matched with the catalogue (5 out of 7).

Minimum working example below

import os
import requests
import numpy as np
from astropy.coordinates import SkyCoord
import calviacat as cvc

stars_a = SkyCoord(np.array([[ 51.59972163, -33.4187023 ],
[ 51.50927589, -33.34720165],
[ 51.55715711, -33.2792019 ],
[ 51.5297432 , -33.34189481],
[ 51.61213712, -33.21934623],
[ 51.46795353, -33.28833715],
[ 51.45707061, -33.29099188],
[ 51.57198152, -33.27501493],
[ 51.63216919, -33.36443673],
[ 51.57411875, -33.40474173]]), unit='deg')

rc2 = cvc.RefCat2('cat.db', min_matches=1)

if len(rc2.search(stars_a)[0]) < 500:
rc2.fetch_field(stars_a)

objids, distances = rc2.xmatch(stars_a)
print(objids)

stars_b = SkyCoord(np.array([
[ 51.50927589, -33.34720165],
[ 51.55715711, -33.2792019 ],
[ 51.5297432 , -33.34189481],
[ 51.46795353, -33.28833715],
[ 51.45707061, -33.29099188],
[ 51.57198152, -33.27501493],
[ 51.57411875, -33.40474173]]), unit='deg')

objids, distances = rc2.xmatch(stars_b)
print(objids)

Very large fields violate the 0.5 deg limit at MAST. Make a way for the user to query a specific location and radius.

https://archive.stsci.edu/prepds/atlas-refcat2/

https://catalogs.mast.stsci.edu/docs/panstarrs.html