Very Computer

I am using awk in my script, I wanted to assign data from awk to a local
variable on the script. How would I do that. This is my script:

#! /bin/sh

filedate=0
monthname="Jan"
ax=`ls -al|awk '{if ($6 == '$monthname') { print $6, $7}}`

I want to assign the data in $7 to the variable filedate, so it would be
available to perform manipulations from the script, outside the awk. I tried
doing " filedate = $7 " it seems this only assigns it to a variable in the
awk.
Thanks

You can't do that directly: shell variables cannot be changed by
subprocesses.
Basically you have two options:
(1) write the value to a file and read it back from there;
(2) capture stdout from the subprocess and pick the result from there.

A third alternative that's sometimes possible is to reorganize
your script so that you call shell from within awk.

You already have an idea of a way to do the latter:

Quote:> ax=`ls -al|awk '{if ($6 == '$monthname') { print $6, $7}}`

although your quotes are a bit off, what you probably mean is

ax=`ls -al|awk '{if ($6 == "'$monthname'") { print $6, $7}}'`

or simpler

ax=`ls -al|awk '$6 == "'$monthname'" { print $6, $7}'`

Quote:> I want to assign the data in $7 to the variable filedate, so it would be
> available to perform manipulations from the script, outside the awk.

After the above you'd have months and days of month in variable ax,
and you could extract the desired data from there.
You'd have the dates for every such file in there though, and the month
for each is same so it doesn't really sound useful;
a while-read loop would probably be better.
But it depends on what you are trying to achieve. Perhaps you
could explain that a bit?

--
Tapani Tarvainen

Quote:>I am using awk in my script, I wanted to assign data from awk to a local
>variable on the script. How would I do that. This is my script:

>#! /bin/sh

>filedate=0
>monthname="Jan"
>ax=`ls -al|awk '{if ($6 == '$monthname') { print $6, $7}}`

>I want to assign the data in $7 to the variable filedate, so it would be
>available to perform manipulations from the script, outside the awk. I tried
>doing " filedate = $7 " it seems this only assigns it to a variable in the
>awk.

What you have above is close; there you get to shellvar ax contents of both
$6 and $7; note though that the ax might contain multiple lines, if the
directory listing has multiple files matching to the specified month.

One way to get multiple variables from awk back yo shell is to run the
awk within eval statement, like
eval `ls -al|awk '{if ($6 == "'$monthname'") { print "filedate=$7"; } }'`
... this will make the awk print out something that the shell can
recognise as commands, and when those command are passed as arguments
to eval (via the `...` structure), they will get executed by the calling
shell.

It is possible to have the awk program generate even largish pieces of
shell programs to be run via the eval statement -- the use is certainly
not limited to just setting variables. This can be needed at places if
it gets impossible to get plain shell code dynamic enough, but for
reason or another (typically political..) Perl cannot be used.