sed or awk replace match with first five characters-CodePudding

I have to replace the word pünktlich with the time(first five characters) at the beginning of each line. It's not always 14:03. The time could change but keeps the same format.

Having:

14:03,RB 87(57140),Donauwörth,pünktlich,Kein Grund

Needed:

14:03,RB 87(57140),Donauwörth,14:03,Kein Grund

The time stands always at the beginning of each line.

Do you have any idea?

CodePudding user response：

Using GNU sed's capturing groups, let file.txt content be

14:03,RB 87(57140),Donauwörth,pünktlich,Kein Grund

then

sed 's/^\([^,]*\)\(.*\)pünktlich/\1\2\1/' file.txt

gives output

14:03,RB 87(57140),Donauwörth,14:03,Kein Grund

Explanation: I use 2 capturing groups: 1st is from beginning of line (^) to first , i.e. it does contain zero or more (*) not (^) commas (,), 2nd is everything between 1st group and pünktlich. I replace what was matched by content of 1st group, content of 2nd group, content of 1st group again. Disclaimer: if any line contains more than one pünktlich then only last one will be changed to time.

(tested in GNU sed 4.2.2)

CodePudding user response：

You can do this simply using awk. let's say your file is named test.csv:

awk 'BEGIN{FS=","; OFS=","}{print $1,$2,$3,$1,$5}' test.csv

Output is:

14:03,RB 87(57140),Donauwörth,14:03,Kein Grund

Explanation: You define input field seperator with FS and output field separator with OFS. Then you set the columns in the order you want. $1 means the first column, $2 means the second column and this is the same for other columns.