My questions is the same as
Command line CSV viewer?
and the column -s, -t command answer is almost perfect - except it doesn't seem to handle empty csv fields as I'd expect. eg given input
col1,,col3
,col2,
produces:
col1 col3
col2
but I would like:
col1 col3
col2
Is there an option to achieve the above with column command or an alternative way to achieve this? (cygwin environment)
-n option exists in debian. e.g.:
>$ echo -e "col1,,col3\n,col2," | column -n -s, -t
col1 col3
col2
It's the furthest thing from elegant, and it's probably something you've already thought of and are looking for a better solution, but I work around this annoyance by doing a series of sed replacements to put a whitespace in empty fields. I have these as functions in my bashrc...
csvcolumn() { sed -e "s/^$2/ $2/" -e "s/$2$/$2 /" -e "s/$2$2/$2 $2/g" -e "s/$2$2/$2 $2/g" $1 | column -t -s$2 ; }
csvcomma() { sed -e 's/^,/ ,/' -e 's/,$/, /' -e 's/,,/, ,/g' -e 's/,,/, ,/g' $1 | column -t -s, ; }
The first one takes two args to be able to specify the delimiter character. The second is the same thing but it only takes one arg and assumes the delimiter is a comma since that's most often what I use anyway.
csvcolumn input.csv ,
or
csvcomma input.csv
Use cat and sed.
cat filename | sed 's/[~\],/\t/g'
