gawk string sorting

staticd · 01-17-2010, 01:55 PM

I have a text file with letters that are all CAP and have no space in between them. I wanted to write a gawk script that pulls out only the letters that are in the "string" and prints them to a new file. After much strife with the asort() function, which produced unpredictable results (gawk version 3.1.3, supposedly this is fixed in version 3.1.7), I wrote a script that gives me reliable and consistent results, so I wanted to share.

Here's the pseudo-code:

Code:

awk 'BEGIN
{FS = "";
RS = "\n"};
{for(i=65;i<=90;i++)
{
letters[i]=sprintf("%c",i)
};
for(j=65;j<=90;j++)
{
match_letters[j]=0
};
for(k=1;k<=NF;k++)
{
for(m=65;m<=90;m++)
{
if($k ~ letters[m])
{
match_letters[m]++
}
}
};
for(n=65;n<=90;n++)
{
if(match_letters[n] > 0)
{
printf"%c",n
}
};
printf"\n"
}' input_file > output_file

The idea is that there is a string with no spaces and you only want to know what letters are in the string and then print them in order. The major problem with awk is that you cannot evaluate characters as an integer.

Jerry Mcguire · 01-19-2010, 02:44 AM

Did you just want to count the stat of appearances of some characters in a file?

% cat a.awk
BEGIN {
FS = ""
RS = "\n"
}

{
for (i = 1; i <= NF; i++) letter_cnt[$i] ++;
}

END {
n = length(your_str)
for(i = 1; i <= n; i++) {
ch = substr(your_str,i,1)
if (ch in letter_cnt) print "'" ch "' : " letter_cnt[ch]
}
}

% gawk -vyour_str="ABC" -f a.awk inputfile

Jerry Mcguire · 01-19-2010, 03:06 AM

or ?

BEGIN {
FS = ""
n = length(your_str)
for (i=1; i<=n; i++) target_ch[substr(your_str,i,1)] = 0;
}

{
for (i=1; i<=NF; i++) if ($i in target_ch) printf "%c", $i
printf "\n"
}

konsolebox · 01-19-2010, 06:12 AM

I just found an awk version of a favorite algorithm of mine - quicksort. I learned the algorithm from a book and I ported it to my bash scripts.

http://en.literateprograms.org/Quicksort_%28AWK%29

jschiwal · 01-19-2010, 06:29 AM

Staticd: Your psuedo-code would be more readable if you used indentation.

staticd · 01-19-2010, 08:05 AM

Quote:

Originally Posted by jschiwal

Staticd: Your psuedo-code would be more readable if you used indentation.

Indeed, I used the appropriate indentation but it must have been lost in translation. I tried several times and the indentation tags never held. I think I know what I did wrong now. The indentation tags have to go before and after each line that must be indented?

jschiwal · 01-21-2010, 02:13 AM

Spaces and tabs are preserved in the CODE blocks.

staticd · 01-21-2010, 09:53 AM

Quote:

Originally Posted by jschiwal

Spaces and tabs are preserved in the CODE blocks.

got it.

staticd · 01-21-2010, 09:55 AM

Quote:

Originally Posted by konsolebox

I just found an awk version of a favorite algorithm of mine - quicksort. I learned the algorithm from a book and I ported it to my bash scripts.

http://en.literateprograms.org/Quicksort_%28AWK%29

Thank you! I will have to see if I can work this into my script...

staticd · 02-10-2010, 08:06 AM

Quote:

Originally Posted by jschiwal

Staticd: Your psuedo-code would be more readable if you used indentation.

Took me a while, but I got it. Thanks for being critical...