I have a tab-separated file (the columns are rsID CHROM POS GENOTYPE) generated by Diganomics with Illumina GSAMD (v2.2) from male DNA, and there are a few hundred loci within X or Y chromosome that contain two alleles, even heterozygous, e.g.
[cut]
rs28463388 X 191998 TT
rs28669107 X 195014 GC
rs6655397 X 201935 GG
rs7890186 X 202458 TG
[skipped]
rs1016793029 Y 8619592 GA
[skipped]
rs527288580 Y 8771750 GA
[skipped]
rs917158084 Y 14871973 TG
[…etc]
There are 1270 such double alleles (7 in Y chromosome and 1163 in the X chromosome) from a total of 734162 number of lines, or 0.17%. What is remarkable is that these 1163 doubles within the X chromosome tend to be clustered, i.e. come consequently one after another mostly, but the 7 in the Y were sparse single lines.