shell - Append a text after a comma in a csv file using Linux bash in a loop - Stack Overflow

admin管理员组

文章数量:1355665

I need to modify a csv file using bash.

Input (a csv file):

firstletter="s"
surname="houston"
emaildomain"@zzz"
input=$(cat 1.csv)

1.csv:

1,1,Susan houston,Director of Services,,
2,1,Christina Gonzalez,Director,,
3,2,Brenda brown,"Director, Second Career Services",,

How can I add a text between the last 2 commas using Linux bash? I tried something like:

for i in $(cat $input);do
        sed -i "s/,$/${firstletter}${surname}${emaildomain},/g" $i;
done

However, that results in an error: sed: -e expression #1, char 5: unterminated `s' command

Expected output:

1,1,Susan houston,Director of Services,[email protected],
2,1,Christina Gonzalez,Director,[email protected],
3,2,Brenda brown,"Director, Second Career Services",[email protected],

I need to modify a csv file using bash.

Input (a csv file):

firstletter="s"
surname="houston"
emaildomain"@zzz"
input=$(cat 1.csv)

1.csv:

1,1,Susan houston,Director of Services,,
2,1,Christina Gonzalez,Director,,
3,2,Brenda brown,"Director, Second Career Services",,

How can I add a text between the last 2 commas using Linux bash? I tried something like:

for i in $(cat $input);do
        sed -i "s/,$/${firstletter}${surname}${emaildomain},/g" $i;
done

However, that results in an error: sed: -e expression #1, char 5: unterminated `s' command

Expected output:

1,1,Susan houston,Director of Services,[email protected],
2,1,Christina Gonzalez,Director,[email protected],
3,2,Brenda brown,"Director, Second Career Services",[email protected],

• bash
• shell

Share Improve this question Follow edited Mar 29 at 2:50 John Kugelman 362k6969 gold badges552552 silver badges596596 bronze badges asked Mar 28 at 16:40 NetRangerNetRanger 5144 bronze badges 10

• There's a range of things wrong with this, first and foremost you can't use sed to replace variable content in-place. Secondly sed is a line-editor used to operate on any number of lines in a file, using bash to feed lines to sed is an anti-pattern. Thirdly all the variables in your replacement string are undefined. You should really feed your script (don't omit the shebang-line) through shellcheck. – tink Commented Mar 28 at 16:44
• @tink if sed can't be used for that, what can do it? Aside from using other tools/language. – NetRanger Commented Mar 28 at 16:53
• 1 @tink I just figured out what you meant, thanks a lot. – NetRanger Commented Mar 28 at 16:59
• I'm pleased to hear that. – tink Commented Mar 28 at 17:08
• 1 It is notoriously hazardous to process CSV files using bash, because if any field is quoted and contains commas, it's hard to process those fields, and there a lot of corner cases (e.g. quoted quotes). It will work for carefully-controlled input (e.g. that outputs no quotes or comma-containing fields), but if your CSV is generated by standard off-the-shelf tools for generic data, you are better off using a tool designed for CSV, such as csvtool (designed for use in bash scripts) or Python (csv module). – Vercingatorix Commented Mar 28 at 17:46

 |  Show 5 more comments

3 Answers 3

Sorted by: Reset to default 5

The question isn't clear but I think this might be what you're trying to do, using GNU awk for "inplace" editing and gensub():

$ cat 1.csv
1,1,Susan houston,Director of Services,,
2,1,Christina Gonzalez,Director,,
3,2,Brenda brown,"Director, Second Career Services",,

$ awk -i inplace 'BEGIN{FS=OFS=","} {$(NF-1)=tolower(gensub(/(.).* (.*)/,"\\1\\2",1,$3) "@zzz")} 1' 1.csv

$ cat 1.csv
1,1,Susan houston,Director of Services,[email protected],
2,1,Christina Gonzalez,Director,[email protected],
3,2,Brenda brown,"Director, Second Career Services",[email protected],

See What's the most robust way to efficiently parse CSV using awk? for more information on processing CSVs with awk.

I'm sure there's some cleverness with sed that would achieve what you want, I'd personally go w/ GNU awk here.

cat a.csv
1,1,Susan houston,Director of Services,,
2,1,Christina Gonzalez,Director,,

gawk -i inplace   'BEGIN{FS=OFS=","}{fn=gensub(/(.).*/,"\\1","1",$3);split($3,ln," ");$5=fn"."ln[length(ln)]"@zzz";print}' a.csv

cat a.csv
1,1,Susan houston,Director of Services,[email protected],
2,1,Christina Gonzalez,Director,[email protected],

-i inplace is a GNU extension that allows awk to emulate seds -i.

The BEGIN section tells awk that both the input and output field separator are commas.

fn=gensub(...) pulls the first letter of the first name (full name being the third field, $3).

We then split the name into an array ln (assuming some people may have middle names).

We set the 5th field (the empty space between the last commas) to the the first letter, and the last element of the array followed by @zzz.

If $5 is not empty:

gawk -i inplace   'BEGIN{FS=OFS=","}$5==""{fn=gensub(/(.).*/,"\\1","1",$3);split($3,ln,"");$5=fn"."ln[length(ln)]"@zzz"}{print}' a.csv

Solution using jq:
With -R parameter (--raw-input) it reads plain text.
With -r parameter (--raw-output) it outputs plain text.

$ jq -Rr --arg emaildomain "@zzz" '
    # split instruction creates an array
    split(",")
    # using JSON object makes the code readable
    |{input_array: .}
    # extract first letter and surname
    |.name = .input_array[2]
    |.firstletter =. name[0:1]
    |.surname = (.name |split(" ") |.[1])
    # construct email
    |.email = (.firstletter+.surname+$emaildomain |ascii_downcase)
    |.output_array =. input_array
    |.output_array[-2] = .email
    |.output_array
    |join(",")
' 1.csv
1,1,Susan houston,Director of Services,[email protected],
2,1,Christina Gonzalez,Director,[email protected],
3,2,Brenda brown,"Director, Second Career Services",[email protected],

本文标签： shellAppend a text after a comma in a csv file using Linux bash in a loopStack Overflow