awk questions

will.flanagan · 04-16-2009, 05:04 PM

Hi LinuxQuestions gurus,

I am trying to learn the awk command for data table manipulation. It seems like a very useful command and I am starting to understand it, but it is still rather opaque to me...

How can one make the field delimiter null so that each space is a field? Perhaps to change:

abcd

to

cbad

Also, are there easy ways to change all of the fields at once? An example might be adding commas:

blah blah blah blah ... blah blah

to

blah, blah, blah, blah, ... blah, blah

Last question: are there shorthand ways of only taking out one field? This would be easier than specifying each field in the print besides the on I'm taking out.
"awk '{print $1 $2 $4 $5 $6....}' example.txt"
(only missing $3 in order to remove it...)
An example might be:

abcde

to

abde

Sorry to bombard the forum with awk questions, but this is very helpful for me... Cheers!

Will

Tinkster · 04-16-2009, 05:20 PM

Quote:

Originally Posted by will.flanagan

Hi LinuxQuestions gurus,

I am trying to learn the awk command for data table manipulation. It seems like a very useful command and I am starting to understand it, but it is still rather opaque to me...

How can one make the field delimiter null so that each space is a field? Perhaps to change:

abcd

to

cbad

For that kind of thing you'd have to use the
FIELDWIDTHS variable .... BEGIN{FIELDWIDTHS="1 1 1 1 1 1 ...."

Quote:

Originally Posted by will.flanagan

Last question: are there shorthand ways of only taking out one field? This would be easier than specifying each field in the print besides the on I'm taking out.
"awk '{print $1 $2 $4 $5 $6....}' example.txt"
(only missing $3 in order to remove it...)
An example might be:

abcde

to

abde

Sorry, no, not that I'm aware of ... a bit of a kludge
and somewhat slower you could work with a loop.

Code:

for(i=1;i<=NF;i++){if(i!=3){printf $i};printf"\n"}

Quote:

Originally Posted by will.flanagan

Also, are there easy ways to change all of the fields at once? An example might be adding commas:

blah blah blah blah ... blah blah

to

blah, blah, blah, blah, ... blah, blah

Sorry to bombard the forum with awk questions, but this is very helpful for me... Cheers!

Will

OFS=", "

Code:

echo blah blah blah blah|awk 'BEGIN{OFS=", "}{$1=$1;print $0}'
blah, blah, blah, blah

Cheers,
Tink

ghostdog74 · 04-16-2009, 07:58 PM

Quote:

Originally Posted by will.flanagan

How can one make the field delimiter null so that each space is a field? Perhaps to change:

abcd

to

cbad

set FS to "".

Code:

# echo abcd | awk 'BEGIN{FS=""}{print $3$2$1$4}'
cbad

Quote:

Also, are there easy ways to change all of the fields at once? An example might be adding commas:

blah blah blah blah ... blah blah

to

blah, blah, blah, blah, ... blah, blah

change OFS to ",".

Code:

# echo "blah blah blah blah ... blah blah" | awk '{$1=$1}1' OFS=","
blah,blah,blah,blah,...,blah,blah

Quote:

Last question: are there shorthand ways of only taking out one field? This would be easier than specifying each field in the print besides the on I'm taking out.
"awk '{print $1 $2 $4 $5 $6....}' example.txt"
(only missing $3 in order to remove it...)
An example might be:

abcde

to

abde

set which field to "". eg $1=$2="" means fields 1 and 2 set to "".

Code:

 # echo "abcde" | awk 'BEGIN{FS=""}{$3=""}1' OFS=""
abde

will.flanagan · 04-17-2009, 10:40 AM

Thanks ghostdog74!,

Very helpful.

Cheers,
Will

PTrenholme · 04-17-2009, 10:59 AM

If you're using gawk (to which awk is often symlinked), see if you have the gawk info file installed. It's quite helpful, and includes many worked out examples.