I have some bash script where I get values of variable, that I would like use in awk. Does it possible include whole awk (like it possible with bash script files) file in bash e.g.:
#!/bin/sh var1=$1 source myawk.sh and myawk.sh:
print $1; 
4 Answers
Bash and awk are different languages, each with their own interpreter of the same name. The tiny sample you show is stripped down too far to make much sense:
- You've marked both files as shell scripts; one using the shebang
#!/bin/shand the other using the extension.sh. Obviously the shell can read shell script, and the command to do so is called.in Bourne shell (orsourcein csh and bash). - The shell script assigns a variable, but you're not using it anywhere. Did you mean passing it on to the awk script?
- Both the awk and shell script use $1, which has different meanings for them (in bash, it's from the command line or a
setcommand; in awk, it's from a parsed input line).
The two tools are often used in tandem, as the shell is better at combining separate programs and awk is better at reformatting tabular or structured text. It was so common that a whole language evolved to combine the tasks; Perl's roots are as a combination of shell, awk and sed.
If you just wanted to pass a variable from the shell script into an awk script, use -v. The man page is your friend.
5 Comments
user710818
. only means for current directory, you can run any script/program using absolute path
Yann Vernier
Don't confuse the command . with the directory .; neither actually means current directory! You'll find the command . accepts any path, and the directory . exists in any directory; but pathnames that don't start with a / will start from the current directory, so . actually refers to the current directory from within itself.
user710818
Dot . always means current directory, if you run ls -la: you will receive: . for current directory, and .. for the up directory. This is simple observation.
Yann Vernier
No, . means the containing directory. Try
ls -di . $PWD .. ../. /. / and compare the inode numbers. There's a . everywhere.Yann Vernier
As an aside, MS-DOS doesn't actually have Unix-style . or .. but emulates them; it also has the notion of a current working directory per device. Things look quite a bit different then.
first of all, if you're writing bash don't use #!/bin/sh that will put you in compatibility mode which is only necessarly if you're writing for portability (and then you have to adhere to the POSIX normative).
now regarding your question you just have to run awk from inside your bash script, like this:
#!/bin/bash var1=$1 awk -f myawk.sh also you should use .awk as extension I guess.
Comments
Or, many ppl do sth like this:
#!/bin/env bash #Bash things start ... var1=$1 #Bash things stop #Awk things start, #may use pipes or variable to interact with bash awk -v V1=var1 ' #AWK program, can even include awk scripts here. ' #Bash things I suggest this page here by Bruce Barnett:
http://www.grymoire.com/Unix/Awk.html#uh-3
You can also use double quote to make use of shell's extract feature but it is confusing.
Personally I just try to avoid those fancy gnu additions of bash or awk and make my scripts ksh+(n)awk compatible.
Comments
As an hardcore AWK user, I soon realized that doing the following was really a huge help :
- Defining and exporting an AWK_REPO variable in my bashrc
#Content of bashrc
export AWK_REPO=~/bin/AWK - Storing there every AWK script I write using the
.awkextension. - You can then call it from anywhere like this :
awk -f $AWK_REPO/myScript.awk $file
or even, using Shebangs and adding AWK_REPO to PATH (withexport PATH=${AWK_REPO}:${PATH})
myScript.awk $file
bashandawk? It's called Perl :P Seriously now, I am not too sure what you're asking... What are you trying to do?