I have a text file which I want to filter using awk. The text file looks like this:
foo 1
bar 2
bar 0.3
bar 100
qux 1033
I want to filter those files with awk inside a bash script.
#!/bin/bash
#input file
input=myfile.txt
# I need to pass this as parameter
# cos later I want to make it more general like
# coltype=$1
col1type="foo"
#Filters
awk '$2>0 && $1==$col1type' $input
But somehow it failed. What's the right way to do it?
pass it in using -v option of awk. that way, you separate out awk variables and shell variables. Its neater also without extra quoting.
#!/bin/bash
#input file
input=myfile.txt
# I need to pass this as parameter
# cos later I want to make it more general like
# coltype=$1
col1type="foo"
#Filters
awk -vcoltype="$col1type" '$2>0 && $1==col1type' $input
parameter=value without the '-v' option. It is better to be explicit - use the '-v' unless there's a problem with it on your system.
parameter=value syntax (without the -v prefix), it has to go between the '$2>0 ...' and the $input arguments. But I doubt there are many systems out there anymore that don't accept -v, and it is best to use it when possible. Also, this method is a bit more robust that John Kugelman's answer: what if coltype held the value xyz{next} {print "garbage"}?"double quote single quote"
awk '{print "'$1'"}'
example:
$./a.sh arg1
arg1
$cat a.sh
echo "test" | awk '{print "'$1'"}'
linux tested
You need double quotes to allow variable interpolation, which means you then need to escape the other dollar signs with backslashes so $1 and $2 are not interpolated. Also you need double quotes around "$col1type".
awk "\$2>0 && \$1==\"$col1type\""
Single quotes inhibit variable expansion in bash:
awk '$2>0 && $1=='"$col1type"
