Inconsistancey between gia and bedtools for unionbedg

[mrvollger]

Hi @noamteyssier,

Thanks so much for adding unionbedg for me! It has been really useful.

However, I have found a corner case where the tool is not in sync with bedtools. Mainly, bedtools allows for empty input files, which sounds odd but is actually really useful when running big pipelines with lots of bedfiles that are sometimes legitimately empty (e.g. ATAC peaks on a very small chromosome).

So e.g. with bedtools I can do:

bedtools unionbedg -i <(printf "") <(printf "chrom1\t0\t1000\t10\n")

And still get the expected output, just with all zeros for the first file.

chrom1  0       1000    0       10

But with gia I get an error due to the empty input:

gia unionbedg -s -i <(printf "") <(printf "chrom1\t0\t1000\t10\n")
Error: Empty file or stream or buffer

I think, in general, not erroring on empty input is actually really valuable with bed files, but in this case in particular it would be extra nice if it did not.

Thanks in advance!
Mitchell

gia --version
gia 0.2.23

[noamteyssier]

Ah I didn't think this would be a useful feature - I'll look into this but should be an easy change to make.