Special input breaks `toolshed.g2.bx.psu.edu/repos/bgruening/split_file_on_column/tp_split_on_column/0.6`

[kysrpex]

usegalaxy.org got an error report on Sentry produced by this tool with very specific input that, despite being apparently valid, breaks it. Please have a look at the Galaxy issue galaxyproject/galaxy#20294, there you will find a link to a Sentry comment with the problematic input.

[bgruening]

From @kysrpex ... not sure if this case is worth to fix it.

 cat problematic_file.txt 
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	.tabular	hg38_knownGeneV46	CDS	57955200	57955269	0.000000	-	1
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57960770	57960844	0.000000	-	1
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57961137	57961294	0.000000	-	0
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57962012	57962176	0.000000	-	0
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57963885	57964069	0.000000	-	2
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57965725	57965827	0.000000	-	0
gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; 	chr15	hg38_knownGeneV46	CDS	57992705	57992818	0.000000	-	0
$ ipython
Python 3.13.3 (main, Apr  9 2025, 07:44:25) [GCC 14.2.1 20250207]
Type 'copyright', 'credits' or 'license' for more information
IPython 9.2.0 -- An enhanced Interactive Python. Type '?' for help.
Tip: The `%timeit` magic has a `-o` flag, which returns the results, making it easy to plot. See `%timeit?`.

In [1]: open("problematic_file.txt", "r").read()
Out[1]: 'gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \t.tabular\thg38_knownGeneV46\tCDS\t57955200\t57955269\t0.000000\t-\t1\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57960770\t57960844\t0.000000\t-\t1\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57961137\t57961294\t0.000000\t-\t0\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57962012\t57962176\t0.000000\t-\t0\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57963885\t57964069\t0.000000\t-\t2\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57965725\t57965827\t0.000000\t-\t0\ngene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; \tchr15\thg38_knownGeneV46\tCDS\t57992705\t57992818\t0.000000\t-\t0\n'

The tool runs awk -F'\t' '{print >> "tmp_out/"\$$column".$infile.ext" }' '$infile'. In plain English, it prints each line to a file named after the value of the column. With that file, the name of the first column is "/corral4/main/jobs/067/731/67731208/working/tmp_out/gene_id "ENST00000537372.5"; transcript_id "ENST00000537372.5"; .tabular" and that messes up everything (specifically .tabular, if it was chr45 it would just work fine).