Finding a string in a text file and posting part of the line

Shell Programming and Scripting
1

What would be the most succinct way of doing this (preferably in 1 line, maybe 2):
searching the first 10 characters of every line in a text file for a specific string, and if it was found, print out characters 11-20 of the line on which the string was found.
In this case, it's known that there are no duplicates in the file, so it either finds the string on one line or none of them.

I'm guessing one of grep, sed or awk can be used but I can't figure out the best way.

2 
sed -n 's/^helloworld//p' file

 sed -n '/^hello/s/..........//p' file
3

maybe like this

# cat 1
thisiscomm11thisisthisis
thisiscomm22thisisthisis
thisiscomm33thisisthisis

# sed 's/^thisiscomm//g' 1
11thisisthisis
22thisisthisis
33thisisthisis
4

Try this also using awk:

awk '{if (substr($0,1,10) == "helloworld") {print substr($0,11)} else {print substr($0,1)}}' file
5 
awk -v v=string 'v==substr($0,1,10){print substr($0,11,10)}' infile
6

Correct one is :

awk '{if (substr($0,1,10) == "helloworld") {print substr($0,11,20)} else {print substr($0,1)}}' file
7

Jairaj

  1. Always read the OP requirement.
  2. Please use [code] tags when you post code.
  3. Don't double-post !
  4. Your solution is wrong , check the OP.
8

Hi, danmero:

I agree. Jairaj's proposed solution is definitely incorrect. However, without some clarification from the original poster, yours could be as well.

If the string being searched must occur in the first ten characters, but can itself be less than 10 characters, it could then occur at different locations within the first 10 characters. If this scenario is a possibility, then

v==substr($0,1,10)

would be insufficient.

If this is the case, then the following would be the correct approach:

awk -v v=string 'i=index($0,v){if (i+length(v)-1<=10) print substr($0,11,10)}'

Again, I think the original poster's problem statement is a bit too vague, leaving some wiggle room for this possibility.

Regards,
Alister

9

This is not about right vs. wrong OR what if suppositions.
The OP ask to "searching the first 10 characters" and that's a fact and we should start from here. Everything else are only suppositions.
I can provide answers/solutions(short) base on facts, otherwise the list will be too long :wink:

Regards,
�

10

Your solution is incorrect, regardless (as was mine). The problem requested printing characters 11-20 if a match is found; that second substr call returns 11-30 (the third argument is length, not an index). The 20 should be a 10.

Regards,
Alister

11

That's correct, I fix my original post :wink:

12

Well, Perl scripts are known to be succinct... :wink:
For the dummy file below, assuming the search string is the 10 character "helloworld" -

$
$ cat f2
helloworld>abcdefghijklmnopqrstuvwxyz
holamundo>abcdefghijklmnopqrstuvwxyz
helloworld>ABCDEFGHIJKLMNOPQRSTUVWXYZ
helloworld>0123456789
bonjourmonde>abcdefghijklmnopqrstuvwxyz
hallowelt>ABCDEFGHIJKLMNOPQRSTUVWXYZ
$
$
$ perl -lne 'print $1 if /^helloworld(.{10})/' f2
>abcdefghi
>ABCDEFGHI
>012345678
$
$

tyler_durden

13

what is wrong Mr danmero?

14 
awk '{if (substr($0,1,10) == "helloworld") {print substr($0,11,10)}}'

You don't want to print the record if no match(your else). My mistake , I take the 20 from you :rolleyes: see alister comments above.

To simplify we can write

awk 'substr($0,1,10) == "helloworld" {print substr($0,11,10)}'