I have CSV file like this:
name_var;type_var;
id;string;
dt_depot;string;
num_artc;string
I want to delete the first's ; and replace the last ones by , so the output would look like:
name_var type_var,
id string,
dt_depot string,
num_artc string
Simply with sed substitution:
$ sed 's/;$/,/; s/;/ /g' file
name_var type_var,
id string,
dt_depot string,
num_artc string
As for awk solution, if input file is supposed to contain only 2 filled columns - you may go with the following awk expression:
$ awk -F";" '{ if (NF==3) $2=$2","; print $1,$2 }' file
name_var type_var,
id string,
dt_depot string,
num_artc string
$ awk -F';' '{sub(/;$/,","); $1=$1} 1' file
name_var type_var,
id string,
dt_depot string,
num_artc string
The script changes every ; at the end of a line to , then converts every remaining ; to , by assigning a value to a field which causes all input field separators (; as set by -F';') to be changed to the output field separator (a blank by default). It then prints the current line by specifying a true condition (1) which invokes the default action (print $0).
With awk something like can do the work:
awk -F\; '{print $1" "$2","}'
If you do not like the last comma it can be rewrited like this:
awk -F\; '{ if (NF != 2) print $1" "$2","; else print $1" "$2;}'
" " between $1 and $2 can be replaced by a comma (space is the default output field delimiter): print $1, $2 ",". This (and your code) would however place a comma on each line.
With this command:
awk '{sub(/;$/,",")sub(/;/," ")}1' file
You get this output:
name_var type_var,
id string,
dt_depot string,
num_artc string
