Error parsing sample names following reheader

[etnite]

I have encountered an error when using bcftools reheader to update sample names using a two-column text file. I have attached an example using the header of the file I'm working with. I have simply changed the extension from .vcf to .txt so that GitHub would allow upload. This file has many samples, but only a few thousand variants.

bcftools query -l in.vcf gives the expected output of sample names (in this case there are 12,833 of them).

Running bcftools reheader in.vcf -s samps_rename_key.txt -o out.vcf completes without any output to stderr.

However, attempting to use the output file, for instance bcftools view out.vcf gives the error:

[E::bcf_hdr_add_sample_len] Duplicated sample name '[Thousands of whitespace-delimited sample names here]
'
Failed to read from out.vcf: could not parse header

I thought this might be a duplicate of #1408, but I have double-checked that all files involved are ASCII encoded, and I have not found any whitespace, tabs, or other unusual characters in any sample names.

in.txt
samps_rename_key.txt

[etnite]

As an update, I tried the substitution in Julia using a dictionary lookup/replace and ended up with the same result.

It strikes me now that many of the sample names being replaced are valid hexadecimal numbers, e.g. "02C01". Not sure if that is significant in any way.

[etnite]

A further update - it appears that the name substitution ends up creating some duplicate sample names. However, it appears that all sample names after the first duplicated one (OH07-159-62) get printed out as one sample name with tabs in the error message. It would be helpful if bcftools reheader could catch the newly-created duplicates, throw an error, and avoid creating a malformed output file in the first place.

[valeriuo]

The problem is your VCF file already contains OH07-159-62 as an original sample name, while your sample name translation file (samps_rename_key.txt) contains the mapping GSTP-69 -> OH07-159-62. Thus, after the sample name translation performed by reheader, your output VCF will contain both the original OH07-159-62 sample name, as well as the translation of GSTP-69.

[etnite]

Thank you @valeriuo