[bash] Read file line by line and split on whitespace

tskuzzy · 07-06-2009, 02:53 PM

I want to read in lines in the file and extract data from each line and store them into variables.

So suppose the input.txt file is:

Code:

So somehow I need to loop through this text file line by line and put the first number in a variable and the second number in another variable.

Is there a simple way to perform this task?

0.o · 07-06-2009, 02:59 PM

cat test | while read one two; do echo $one -- $two; done

acid_kewpie · 07-06-2009, 03:02 PM

while read LINE
do
var1=$(echo $LINE | cut -f1)
var2=$(echo $LINE | cut -f2)
do_something_sexy
done < input.txt

that's just with bash though (or rather bash and GNU tools - it's posisble to replace the echo / cut with something like var1=$(LINE%#\ ) but I can never remember the right syntax for the real short way so i always end up echoing)

[hands up that i didn't know read could go to multiple variables, props to mr 0.o]

better off using awk though

awk '{var1 = $1; var2 = $2; sexy_time}' -f input.txt

pixellany · 07-06-2009, 03:17 PM

Yes!!

Set up an array for each column---let's say "this" and "that". On the first pass, you'll assign values to "this[1]" and "that[1]".

a straightforward way to assign variables based on columns is with AWK---eg:
varname = awk '{print $1}' filename

So, you might have something like this (some is pseudo-code, not tested):

Code:

n=1
while read line; do
var1 [n] = $(echo $line | awk '{print $1}')
var2 [n] = $(echo $line | awk '{print $2}')
increment n
done
printf to print var1 and var2 in desired format

tskuzzy · 07-06-2009, 03:24 PM

Wow, nice variety of solutions!
I'm using 0.o's solution since its the simplest and most elegant one for my particular problem.

awk and cut are still way over my head to use effectively so I try to stay away from them :P
One of these days I'll learn...

Thanks for all the help