Shell Programming and Scripting

In AWK script how do I store data in variables for later use.
I have a multiline input and I do not want to print the data read on the console
Thnaks in advance.
Nilotpal.

You can pipe the results through read, e.g. to get the first and second fields from the third line of a file....

awk 'NR==3{print $1, $2}' file1 | read var1 var2

Hi Ygor,
I created the following file <<t1.awk>>
BEGIN{FS = "|"; RS = "\n"}
NR==2{print $1, $2} | read var1 var2
echo $var1 ' ' $var2

as my data is "|" delimited
then from $prompt I am writing:
awk -f t1.awk abc.flag (my source file)
it gives me an error:
awk: syntax error at source line 2 source file t1.awk
context is
NR==2{print $1, $2} >>> | <<< read var1 var2
awk: bailing out at source line 4

Please help
Thanks in advance

The read command is a unix shell command NOT an awk command

$ cat abc.flag
1|2|3
e|g|3
1|2|3
$ cat t1.awk
BEGIN{FS = "|"; RS = "\n"}
NR==2{print $1, $2}
$ awk -f t1.awk abc.flag | read var1 var2
$ echo $var1 ' ' $var2
e g

Hi Ygor,
Thanks for your prompt help.
Can we read the files form line 1 and 2 at one attempt through awk?
individually I am able to read the lines and redirect to variables.
Can I read both the lines together and redirect to 6 variables (I am refering to the file layout that you gave as example:
a|b|c
1|2|3
x|o|p

my t1.awk now looks as follows:
BEGIN{FS = "|"; RS = "\n"}
NR==2{print $1, $2 $3}
NR==1{print $1, $2}

then I wrote:
awk -f t1.awk abc.flag | read - and then I got confused as the $1 and $2 variables will overwrite the previous $1 and $2 variables.
How do I retain the values that I have read in the first line - (NR==2{print $1, $2 $3})
and then go over to the next line.

Hope I cud explain.

Regards,
Nilotpal.

Not sure I understand entirely. Perhaps capture the output from awk in an array?

$ cat abc.flag
a|b|c
1|2|3
x|o|p
$ cat t1.awk
BEGIN{FS = "|"; RS = "\n"}
NR==2{print $1, $2 $3}
NR==1{print $1, $2}
$ set -A var -- `awk -f t1.awk abc.flag`
$ echo ${var[0]} ' ' ${var[1]} ' ' ${var[2]} ' ' ${var[3]}
a b 1 23

Or perhaps assign to variables in awk....

$ cat t1.awk
BEGIN{FS = "|"; RS = "\n"}
NR==1{j=$1; k=$2}
NR==2{print j, k, $1, $2, $3}
$ awk -f t1.awk abc.flag
a b 1 2 3

hi Ygor,
Thanks for the assistance.
I am able to use the local variable approach (j=$1...) but not the array approach. The arracy approach is taking the entire string 'awk -f t1.awk abc.flag' as an argument and not the output of awk awk -f t1.awk abc.flag
My t1.awk now looks like:

BEGIN{FS = "|"; RS = "\n"}
NR==1{j=$1;k=$2}
NR==2{print j, k, $1, $2, $3}

then $ awk -f t1.awk abc.flag
gives me:
a b 1 2 3
on the console.
How can I capture this console output so that I can use them in a shell script to insert into a oracle table?

I guess I cannot do database activities (insert/select) within a AWK script.

My Insert shell script is:
sqlplus $USERNAME/$PASSWD@clbd << END
insert into nr_test (COL1, COL2, COL3, COL4, COL5 ) VALUES ('$var1', '$var2', $var3, $var4, '$var5);
EXIT;
END

so typically i need the data in the awk variables j,k,$1, $2, $3 to be acceptted into $var1', '$var2', $var3, $var4, '$var5

Please advice.