script variables within awk

[Sylvain CROUET]

Hi!

I use variables within a script and I would like to use them within an awk command without the -v option.
For example:
THRESHOLD1=90
THRESHOLD2=95
bdf | awk '{print $1, ":$THRESHOLD1:$THRESHOLD2"}' > FILE

I know this would work:
bdf | awk -v thr1=$THRESHOLD1 -v thr2=$THRESHOLD2 '{print $1, ":thr1:thr2"}' > FILE

but I am looking for something... smarter.

Can someone help me?

[ian Dennison]

perform a "print" of the line "bdf | awk '{print $1, $THRESHOLD1:$THRESHOLD2"}' " and redirect to a temporary file. Then run that file using the 'sh' command.

This is helpful if the number of variables to ber passed is not known

Alternatively you can string these variables together to pass only 1 variable to the awk statement!

Share and Enjoy! Ian

[Sylvain CROUET]

I prefer the second solution.

[Rodney Hills]

Using "-v" is the standard awk mechanism for passing variables. You could use "perl" instead.

bdf | perl -nae 'print "$F[0]:$ENV{THRESHOLD1}:$ENV{THRESHOLD2}' >FILE

HTH

-- Rod Hills

[curt larson_1]

depends on which version of awk your running,
if it is a newer version, such as on an 11i system, you can use the ENVIRON array to access the values

export THRESHOLD1=90
export THRESHOLD2=95
bdf | awk '{print "$1, :ENVIRON[THRESHOLD1]:ENVIRON[THRESHOLD2]";}' > FILE


and of course you can also pass values to awk using the command line via argc and argv somewhat like a c program

[Sylvain CROUET]

Interesting ideas but I do not have very recent servers and perl is not installed on all machines.

[curt larson_1]

you'll need double qoutes around the variable name

ENVIRON["THRESHOLD1"]

and you don't have to export the variables;
export THRESHOLD1=90

[Michael Schulte zur Sur]

Hi,

believe Curt, environ is what you need to evaluate envoronment variables from within awk. See man awk.

greetings,

Michael

[Sylvain CROUET]

Yes, ENVIRON seems to be what I need, but my Solaris machines do not seem to implement a correct version of awk as there is no ENVIRON variable within the awk manual page and it does not work as it does under HP-UX.

[Dietmar Konermann]

On Solaris you may try /usr/xpg4/bin/awk instead.

[Michael Schulte zur Sur]

Hi,

can you post an example and why it does not work, i mean any error messages?

greetings,

Michael

[Jean-Luc Oudart]

awk -v would be the solution , I don;t know why you don't want to use it !

Nevertheless, you can run something like :
(echo 90 && echo 95 && bdf) |
awk 'BEGIN{getline; th1=$0; getline; th2=$0; }
{print $1,th1,th2;}'

Regards,
Jean-Luc

[Michael Schulte zur Sur]

Hi,

there is an alternativ to environ. Store the values of the environment variables into a file and then read them ony by one.

getline < extvarfile; var1=$0;
getline < extvarfile; var2=$0;

greetings

Michael

[Sylvain CROUET]

I finally used Ian's solution; but Curt's solution, with the precision brougth by Dietmar, works fine too.

[James Lynch · Accepted solution]

It's all in the quoting. You could do this:

THRESHOLD1=90
THRESHOLD2=95
bdf | awk "{print \$1, \":$THRESHOLD1:$THRESHOLD2\"}" > FILE


JL

[Rory R Hammond]

I am going to show two other suggestions that didn't show up in the thread.

THRESHOLD1=90
THRESHOLD2=95
bdf | awk '{print $1, ":"THRESHOLD1":"THRESHOLD2}' THRESHOLD1=$THRESHOLD1 THRESHOLD2=$THRESHOLD2 > FILE

or

bdf | awk '{print $1, ":"THRESHOLD1":"THRESHOLD2}' THRESHOLD1=90 THRESHOLD2=95 > FILE

Rory