Insert newline when grep result is empty

Shell Programming and Scripting
1

Given a csv file with 40 columns with name, address, hometown etc.
I use a bash command in 1 line which:

  1. gets the address column and pipes that to
  2. grep the first digit and everything that follows

Command:

awk -F ";" '{print $19}' /Users/jb/Desktop/ReorderTempTotal.csv  | grep -o "\d.*"

It generates the results I want, except in the case that the grep result is empty.
When there is no digit present, the command prints nothing and the result will end up with 1 line less. That gives a problem in further processing. I need the same number of lines as in the original file. So in that case I'd like to insert a new line for further processing.
And the code needs to be in one line

2

Hmm, normally an empty line gives more processing problems than no line.
Perhaps you can store in a variable and echo it:

result=$(
awk -F ";" '{print $19}' /Users/jb/Desktop/ReorderTempTotal.csv  | grep -o "\d.*"
)
echo "$result"
3

What MadeInGermany suggested is the best, just store result in variable, meanwhile you can try below one too

[akshay@localhost tmp]$ cat f
2
Haider
Bash

[akshay@localhost tmp]$ touch ff
[akshay@localhost tmp]$ wc -l ff
0 ff

[akshay@localhost tmp]$ cat f |  grep -o "\d.*"  || echo "newline"
der

[akshay@localhost tmp]$ cat ff |  grep -o "\d.*"  || echo "newline"
newline

So your command would be

awk -F ";" '{print $19}' /Users/jb/Desktop/ReorderTempTotal.csv |  grep -o "\d.*"  || echo

Note : grep's default mode is (iirc) POSIX regex, and \d is pcre. You can either pass -P to gnu grep, for perl-like regexps, or use [[:digit:]] instead of \d ( BSD grep's -E mode includes \d )

4

Sadly, both are no working

@MadeInGermany
I need this result with blank lines because the result should match with the original cvs. Otherwise data from different lines get mixed up.
The resulting data wil be used for a new column added to the original

I added the cvs I use for the testing

5

What is expected output for the given sample ?
In case if you expect newline to be printed for the line which does not content digits, you may try below command

awk -F ";" '{print match($19,/[0-9]+.*/)?substr($19,RSTART,RLENGTH):""}' /Users/jb/Desktop/ReorderTempTotal.csv
6

Well.. this works! Almost, because it's inserting 2 lines instead of 1
and when I use

awk -F ";" '{print match($19,/[0-9]+.*/)}' /Users/jb/Desktop/ReorderTempTotal.csv

it inserts a zero
It is promising though :slight_smile:

7

Use this

awk -F ";" '{print match($19,/[0-9]+.*/)?substr($19,RSTART,RLENGTH):""}' /Users/jb/Desktop/ReorderTempTotal.csv
8

Great! This is working!
Thanks to you all guys!