AWK

[Chern Jian Leaw]

HI,

I trying to extract the fields in a txt file separated by commas as below:

abc,hhh,lll,kkk,ppp,
dajd,dkjdk,nnn
mmm

I tried using the script below to extract the these fields, place them into an array and write the array contents into a file.

cat test.sh:

BEGIN {
split($0, fieldArray, ",")
for(i in field){
print field[i]
field[i] >> groupfile
}
}

However, it does not produce any results into the groupfile, but produces a parse error at:
field[i] >> groupfile

I tried running the script as follows on the command line:
$awk -f test.sh inputfile

How should I solve this problem ?

Also, what if I were to have 2 different separators in the input file:
wmt:*abc-nn-mm:306:abc,kkk,ppp,...
....,abn

I would only need to extract the fields separated by the commas, not the colons.

How should also do this?

Thanks.

[Heiner E. Lennackers · Accepted solution]

Hi,

try this:

awk '{
split($0, field, ",");
for (i in field) {
print field[i] >> groupfile
}
}' input_file

In split you can use a regular expression as seperator. if you want to use * and , as separator, just try:
split ($0, field, "[*,]")

Heiner

[Deepak Extross]

You'll find it easier to use the -F option to specify ' as your delimiter. See man awk for details.

[Chern Jian Leaw]

HI Heiner,

It produced an ^Invalid char '`' in expression.

Is the sign after awk, a ` or ' and the sign after closing statement a ' or `, as either one produces the same error...?

Please advice.


Thanks.

[John Carr_2]

Hi

try this

cat filename | akw -F"," '{ print $1 $2 $3 }'

cheers
JOhn.

[Chern Jian Leaw]

HI John,

If the text file has only very little entries, then it works. Unfortunately, the file is rather huge with lots of such entries separated by commas.

Please advice.
Thanks.

[Deepak Extross]

Chern,
These are single-quotes (the ones above the ? key).

The following works fine for me:
awk '{
split($0, field, ",");
for (i in field) {
print field[i]
}
}' input_file >> ./groupfile

[John Carr_2]

Chern

can you attach a section of the file so we know what we are dealing with

cheers
John.

[Robin Wakefield]

Hi,

Try:

awk -F, '{for (i=1;i<=NF;i++){print $i}}' infile > outfile

Rgds, Robin

[H.Merijn Brand (procura]

l1:/tmp 102 > cat >xx
aa,bb,dvfgrh:nfbe,fgr;hfhr-z
243,9474632 jfkweh,fdfwe
l1:/tmp 103 > perl -naF, -e '$"="\n";print"@F"' xx
aa
bb
dvfgrh:nfbe
fgr;hfhr-z
243
9474632 jfkweh
fdfwe
l1:/tmp 104 >