Counting number of character occurrences per line

Question

I have a file that looks like this: ( Note : A*, B*, C* are placeholders). The file is delimited by ;

AAAA;BBBB;CCCCCCCC;DD;EEEEEEEE;FF;
AAA1;BBBBB;CCCC;DD;EEEEEEEE;FFFFF;
AAA3;BB;CCCC;DDDDDDDDD;EEEEEEE;FF;

I m trying to write a small script that counts the number of occurrences of the delimiter ; and if it is lesser or greater than 5, output said line to a text file.

delim=";"

while read line
do  
    n_of_occ=$(grep -o "$delim" <<< "$line" | wc -l)

    if [[ $n_of_occ < 5 ]] || [[ $n_of_occ > 5 ]]
    then
        echo $line >> outfile
    fi
done

For some reason, this doesn't seem to work and my output is garbled. Could someone assist or provide a different way to tackle this? Perhaps with Perl instead of bash?

You should try to supply a properly representative set of data. Every line of your sample has six semicolons `;` which, according to your rules, means they should all be printed. Once you have said *"The file is delimited by `;`"* there's little point in giving an example unless it tests the criteria and is accompanied by your corresponding required output — Borodin, May 17 '16 at 13:27

score 3 · Answer 1 · answered May 17 '16 at 11:25

3

This is ridiculous easy with awk:

awk -F\; 'NF!=6' file > outfile

answered May 17 '16 at 11:25

Juan Diego Godoy Robles

14,447
2
38
52

Kent · Answer 2 · 2016-05-17T11:32:27.603

1

I would take this one-liner:

awk '{x=$0}gsub(";","",x)!=5' file

edited May 17 '16 at 11:32

answered May 17 '16 at 11:26

Kent

189,393
32
233
301

score 1 · Answer 3 · answered May 17 '16 at 11:37

1

Easy in Perl:

perl -ne 'print if tr/;// != 5' input_file > output_file

-n reads the input line by line
the tr operator returns the number of matches

answered May 17 '16 at 11:37

choroba

231,213
25
204
289

Jahid · Answer 4 · 2016-05-17T11:54:43.010

1

With sed you can do this:

sed '/^\([^;]*;\)\{5\}$/d' file > outfile

It deletes the lines with exactly 5 commas (;) and sends the output to outfile.

Or if you want your own code to work, then make the following changes:

replace done with done <file
replace [[ with (( and ]] with )) i.e use ((...)) instead of [[...]]

edited May 17 '16 at 11:54

answered May 17 '16 at 11:45

Jahid

21,542
10
90
108

score 1 · Accepted Answer · answered May 17 '16 at 13:07

Unfortunately every line in your sample data has six semicolons, which means they should all be printed. However, here is a one-line Perl solution

$ perl -ne'print if tr/;// != 5' aaa.csv
AAAA;BBBB;CCCCCCCC;DD;EEEEEEEE;FF;
AAA1;BBBBB;CCCC;DD;EEEEEEEE;FFFFF;
AAA3;BB;CCCC;DDDDDDDDD;EEEEEEE;FF;

Counting number of character occurrences per line

5 Answers5