getting output using awk

[Florinda Adato]

Hi,

I have an awk script that actually parse the fields in the file but my problem is, I must get/grep the field value in another file.
My script looks like this:

awk '{FIELD1=$1; FIELD2=$2;
printf ("%15s %10s", FIELD1, FIELD2)}' FILE1

I wanted to get/grep the FIELD2 from FILE2 in order to get another data from FILE2, how do I get it?

[Jean-Luc Oudart]

Within your awk script you can also tun a grep command:

s1=sprintf(" grep %s FILE2",FIELD2);
system(s1);

Note for each FIELD2 it will generate a process grep FILE2.

Rgds,
Jean-Luc

[Florinda Adato]

Thanks. But how will I save the output of my grep to another variable?

[Rodney Hills]

How about using "join" to combine the fields you want from file1 and file2.

Example-
join -j1 1 -j2 1 -o 1.1,2.2 | awk ...

This would send the common key field for file1 and file2 and the second field of file2 to awk.

"join" works like a SQL join, treating the 2 text files like tables.

HTH

-- Rod Hills

[Florinda Adato]

Thanks.

Join works fine but I need find the matched
of field2 from file2 using awk.

My FILE1 have these fields:
12345 20
12346 30

My FILE2 have these fields:
20 hello
30 thanks

What I want is to get the string "hello" or "thanks" in my awk.

The output will be something like these:
12345 20 hello
12346 30 thanks

[Hoefnix]

try this one(if the files are very big it will take a while because it running through file1 for each entry in file2:

cat file2 | while read indx word
do
indx2=""
cat file1 | while read field1 indx2
do
if [ $indx -eq $indx2 ]
then
echo "$field1 $indx $word"
fi
done
done

file1:
12345 20
12346 30

file2
20 hello
30 thanks


Output:
12345 20 hello
12346 30 thanks

Regards,

Peter

[Florinda Adato]

My file is about 700k lines, I need something faster.

[Jeroen Peereboom]

Are both files large?
If one is small, it may be possible to fill an array with the values and then process each line of the large file, looking up the value in the array.

If both are large, sort both files on the field they have in common. Then read both files, advancing in a file if the common field is less than the field of the current line of the other file.

This may be tricky in awk.

JP.

[Elmar P. Kolkman]

(
cat FILE2
echo --- starting data ---
cat FILE1
) awk '/--- starting data ---/ {infile1=1}
infile1==0 { response[$1]=$2 }
infile1==1 { printf("%15s %10s %s\n",$1,$2,response[$2]);}'

This should do the trick...

[Jeroen Peereboom]

Elamr,

looks nice. Wonder what Florinda thinks (and awards?)

JP.

[Elmar P. Kolkman · Accepted solution]

Jeroen, Nice to see you online... ;-) Long time no see. While I'm back at the company we worked together in the past...

Florinda, if you check my solutions, put a pipe ('|') between the closing bracket ('}') and awk, otherwise it won't work (and complain about a syntax error).

[Florinda Adato]

thank you very much, until next time :-)

[Florinda Adato]

Hello.

How about if my files are something like these:

::::::::::::::
FILE1
::::::::::::::
12345 123 123 20
12346 123 123 30
12347 123 123 20
12348 123 123 90
::::::::::::::
FILE2
::::::::::::::
20 hello
30 thanks


OUTPUT IS:
12345 123 123 20 hello
12346 123 123 30 thanks
12347 123 123 20 hello
12348 123 123 90

Sorry I'm very new in using awk.

Thanks

[Elmar P. Kolkman]

You can use the same script I sent earlier, but replace the line
'infile1==1 { printf("%15s %10s %s\n",$1,$2,response[$2]);}'

with

'infile1==1 { printf("%s %s\n",$0,response[$NF]);}'

This will work, no matter how many fields are in FILE1

[Marc Roger]

The same in perl, and it's probably faster.

#!/usr/contrib/bin/perl

open(FILE2, $ARGV[1]);
while() {
($field1, $field2)=split;
$map{$field1}=$field2;
}
close(FILE2);

open(FILE1, $ARGV[0]);
while() {
chop;
@l=split;
print "$_ $map{$l[$#l]}\n";
}
close(FILE1);