[SOLVED] awk

danielbmartin · 02-05-2012, 08:03 PM

I'm learning awk and wrote a simple test program. The desired operation is to read a file of words and output only those which...
- are five characters long
- have the same letter in position 2 and 5.
A sample input file...

Quote:

apple
orange
onion
banana
war
peace
surrender

The desired output file...

Quote:

onion
peace

I wrote this...

Code:

cat < $InFile \
|awk '{length($0)=5 && substr($0,2,1)==substr($0,5,1)}{print}' \
> Work04

... and get a syntax error. Please advise.

Daniel B. Martin

jhwilliams · 02-05-2012, 09:25 PM

Very close! How about this:

Code:

cat < $InFile | \
awk '{ if (5 == length($0) && substr($0,2,1) == substr($0,5,1)) print $0 }' \
> $OutFile

firstfire · 02-05-2012, 09:42 PM

Hi.

For learning purposes here is another solution

Code:

$ awk -F '' 'NF==5 && $2==$5' infile.txt 
onion
peace

-F '' -- set field separator to empty string, in which case awk treats each character as a field.
NF -- number of fields; in this case -- length of words.
$2, $5 -- second and 5th character in the string.

EDIT:
1) Note that if there are no action "{...}" after a pattern (or logical expression) then, by default, awk assumes '{print $0;}'

2) from info gawk:

Quote:

* Null strings are removed when they occur as part of a non-null
command-line argument, while explicit non-null objects are kept.
For example, to specify that the field separator `FS' should be
set to the null string, use:

awk -F "" 'PROGRAM' FILES # correct

Don't use this:

awk -F"" 'PROGRAM' FILES # wrong!

In the second case, `awk' will attempt to use the text of the
program as the value of `FS', and the first file name as the text
of the program! This results in syntax errors at best, and
confusing behavior at worst.

danielbmartin · 02-06-2012, 02:44 PM

Quote:

Originally Posted by jhwilliams

Very close! How about this:

Code:

cat < $InFile | \
awk '{ if (5 == length($0) && substr($0,2,1) == substr($0,5,1)) print $0 }' \
> $OutFile

Thank you for correcting my faulty code. Your solution is readable and I like it.

I made the next step by parameterizing this line.

Code:

# Parameterize the word length and "must match" character positions.
WL=6
p1=2
p2=5
cat < $InFile \
|awk '{ if (length($0)=='"$WL"'&&substr($0,'"$p1"',1)==substr($0,'"$p2"',1)) print $0 }' \
> $Work05

This works but the combination of single quotes and double quotes detracts from readability. Is there a cleaner way?

Daniel B. Martin

danielbmartin · 02-06-2012, 02:52 PM

Quote:

Originally Posted by firstfire

For learning purposes here is another solution

Code:

$ awk -F '' 'NF==5 && $2==$5' infile.txt 
onion
peace

-F '' -- set field separator to empty string, in which case awk treats each character as a field.
NF -- number of fields; in this case -- length of words.
$2, $5 -- second and 5th character in the string.

Thank you, firstfire, for this remarkably concise solution. Setting the field separator character to the null string is clever!

I made the next step by parameterizing this line.

Code:

# Parameterize the word length and "must match" character positions.
WL=6
p1=2
p2=5
cat < $InFile \
|awk -F '' 'NF=='"$WL"' && $'"$p1"'==$'"$p2"' ' \
> $Work07

This works but the combination of single quotes and double quotes detracts from readability. Is there a cleaner way?

Daniel B. Martin

firstfire · 02-06-2012, 03:11 PM

Hi.

Quote:

Originally Posted by danielbmartin

This works but the combination of single quotes and double quotes detracts from readability. Is there a cleaner way?

Yes, definitely.

Code:

$ awk -v WL=5 -v p1=2 -v p2=5 '{ if (length==WL && substr($0,p1,1)==substr($0,p2,1)) print}' infile.txt 
onion
peace

As you can see, one may pass variables to awk script using `-v' option.

I'd like to dissuade you from writing awk programs in C (as well as C++ programs in C etc)

IMHO, each language has it own preferable thought patterns, and you should learn them, not only the syntax. In case of awk, typically, you should think of input data as a sequence of records, each one consists of fields. Using this simple paradigm you can move mountains!

s4sandeep · 03-14-2012, 12:31 PM

Hi,

I have just started with unix and today started learning awk, this is my first post on this forum and do not know where to start a new thread so posting on this thrad.

Problem Details:
I tried to run the follwing command
ls -l | grep -v total | awk '{ print size is $5 bytes for $2 }' and it showed me this error:

syntax error The source line is 1.
The error context is
{ print size is $5 bytes >>> for <<< $2 }
awk: The statement cannot be correctly parsed.
The source line is 1.

while if I running this command : ls -l | grep -v total | awk '{ print size is $5 }' , it runs successfully with output as:
164
17146-rw-r--r-- 1 goyalank users 164 Mar 7 09:20 email
-rw-r--r-- 1 goyalank users 17146 Mar 7 13:49 task

To mention the detail the output of ls -l is :

-rw-r--r-- 1 goyalank users 164 Mar 7 09:20 email
-rw-r--r-- 1 goyalank users 17146 Mar 7 13:49 task

Any help is appreciable.

firstfire · 03-14-2012, 02:01 PM

Hi.

You should quote literal strings, like this:

Code:

ls -l | grep -v total | awk '{ print "size is "$5" bytes for "$2 }'

s4sandeep · 03-14-2012, 03:42 PM

Thanks a ton! That worked......
Can you also tell the step to create a new thead for a particular problem like the one I faced below....thanks again in advance.

firstfire · 03-14-2012, 10:56 PM

Quote:

Originally Posted by s4sandeep

Thanks a ton! That worked......
Can you also tell the step to create a new thead for a particular problem like the one I faced below....thanks again in advance.

Click on "Forum Tools" (next to "Search This Forum") -> "Post a New Thread".