Option to set a default delimiter

[HerbCSO]

OK, can I just say first of all that I'm IN LOVE with this toolkit!? It is an absolute joy and it fills a gaping void for me! ;]

Anyway, I work with a lot of TSV files (tab-separated) and it's kind of annoying to have to type -d "\t" for every single command I run (but thank you for providing the option!). I'd love to be able to change the default separator from , to \t.

Unfortunately it's not easy to alias since the -d must come AFTER the command. I suppose I could do something like xsvtab() { xsv $1 -d "\t" ${@:2}; } (and maybe even alias xsv=xsvtab if I'm feeling really lazy - but then I lose the ability to override the delimiter if I actually do have a different type of file and I have to run it with \xsv to use the original command instead of the alias (specifying -d again with the above alias results in an Invalid arguments. error because now -d is specified twice)). That all feels a little clunky, although it does kind of work for my use case, it would just be slicker to be able to override the default.

[BurntSushi]

I'd support an environment variable to set this.

One trick you may not be aware of (and it may not work for your use case), but if your csv files have a .tsv extension, then xsv should use a tab delimiter automatically.

[HerbCSO]

Oh, that's cool, I was indeed not aware of that. Unfortunately a lot of files I work with don't have a .tsv extension for... reasons...

But an environment variable for that would be totally awesome!

[ghuls]

Maybe environment variables based on awk variable names would be a good idea:

XSV_FS = input field separator (input field delimiter)
XSV_OFS = output field separator r (output field delimiter)

An environment variable for specifying additional file extensions for automatically interpreting files as TAB-separated files, would be useful too. A lot of bioinformatic related file formats are TAB-separated file formats.

XSV_TSV='bed|gtf|gff|tsv|vcf'

[link2xt]

How about defaulting to automatic separator detection? Separator is either ',' or '\t', whichever comes first.

[BurntSushi]

@ilabdsf Won't work because escaping/quoting permits either of those characters to be present before the first field separator.

[link2xt]

@BurntSushi right, that is why I say "default to". If someone wants a robust script, he can write "-d,". But when just using xsv from the command line, I think it is acceptable. It is very unlikely to have have tabs or commas in column names.

[iliekturtles]

A bash function can be setup to always use a specific delimiter:

function xsvt() {
    local cmd=$1
    shift && command xsv $cmd -d"\t" $@
}

[camerondavison]

I vote for an environmental variable that can set the default input/output for all of the pipes maybe just something like
XSV_DEFAULT_DELIMITER

[ghuls]

@iliekturtles
The bash function should be a a lot more complex than that.

Some things that don't work:

$ xsvt -h
Unknown flag: '-d'

Usage:
    xsv <command> [<args>...]
    xsv [options]

# No output (instead of help):
$ xsvt

# Overriding -d does not work:
$ printf '1\t2\n' | xsvt input -d '\t'
Invalid arguments.

Usage:
    xsv input [options] [<input>]

[camerondavison]

Not sure if you take pull requests or not, but I created one for my proposal. I have not written much rust code, but I thought that this seemed easy enough things to dive and and learn some. I was thinking about trying to document the variable somewhere, but was not really sure where. Do you want me to put it into all of the help strings?

[jimmywan]

@BurntSushi can you take a look at #94 ?

[richardjharris]

alias tsv='xsv -d "\t"'

[nickray]

The main issue for me is that the output delimiter cannot always be configured. If I get some TSV files (with embedded commas), and do something like xsv cat, then since there is no --out-delimiter flag, the files get corrupted.

So to make everything composable, it would be great if an env delimiter influences both input and output.

[BurntSushi]

@nickray You can add xsv fmt -t '\t' to the end of your pipeline.

[nickray]

I didn't realize that xsv cat rows (and probably others) inserts double quotes, I would have expected both.csv to be broken in the following (my mental model of CSV is something like line.split(",")...):

printf "a\tb\nx,y\tz\n" > file1.tsv 
printf "a\tb\nx\ty,z\n" > file2.tsv

echo ":: both.csv"
xsv cat rows -d '\t' file?.tsv > both.csv
cat both.csv

echo ":: both.tsv"
xsv fmt -t '\t' both.csv > both.tsv
cat both.tsv

Good to know it's not!

:: both.csv
a,b
"x,y",z
x,"y,z"
:: both.tsv
a       b
x,y     z
x       y,z

Would still be helpful to avoid the quotation dance, and use tabs (or 0x1f) throughout by just doing something like export XSV_DELIMITER='\t'; xsv cat rows file?.tsv > both.tsv, particularly when chaining xsv calls (or, for instance, xsv partition would seem to need an xargs or parallel call to end up with TSV splits) . Maybe my Rust skills are soon good enough to contribute soon 🌱

[erlangparasu]

Any updates?