Writing a bash script more elegantly

Question

So I have the following script working as a daily workaround until the dev people fix:

#!/bin/sh

sed -i 's/,/;/g' /file_path/filename_date.csv
sed -i 's/NAME;NAME/NAME,NAME/g' /file_path/filename_date.csv

It works but when I look at it I think something is missing or isn't elegant. Maybe is my way of thinking being so much straight forward.

What is your opinion?

What is the script supposed to do? Replace all `,` (except those between `NAME` and `NAME`) with a semicolon? — chepner, Jun 30 '20 at 13:47
Yes. when doing the first one it replaces all ```,``` with ```;``` but at one field ```NAME,NAME``` doing that command changes the ```,``` with ```;``` and I need that field to be with ```,``` if not, importing the csv file gets messed up — anmoreira, Jun 30 '20 at 14:41

score 0 · Accepted Answer · answered Jun 30 '20 at 13:47

0

Like this:

sed -i 's/,/;/g; s/NAME;NAME/NAME,NAME/g' file

or

sed -i -e 's/,/;/g' -e 's/NAME;NAME/NAME,NAME/g' file

answered Jun 30 '20 at 13:47

Gilles Quénot

1

Thank you for replies. I was complicating too much on my head but it is still straight forward your suggestions. – anmoreira Jun 30 '20 at 14:43

score 0 · Answer 2 · answered Jun 30 '20 at 13:49

0

Maybe with extended regular expressions and capture groups:

sed -Ei 'y/,/;/;s/(NAME);\1/\1,\1/g'

The y command is for transliteration.

answered Jun 30 '20 at 13:49

Benjamin W.

2 Answers2