Splitting long lines in bash with numbering

user_4151158 0
Viewed 60

I'm trying to write script that partition very long text (one line) to multiple lines. The requirement is that every line must have prefixes with numbering. I tried using sed but I have problem with numbering matches. My code:

 sed -e "s/.\{4\}/part = &\\
 /g"

Input:

AAAAAAAAABBBBBBBAAAAAAABBBBBBB

Output:

part = AAAA
part = AAAA
part = ABBB
part = BBBB
part = AAAA
part = AAAB
part = BBBB
BB

Expected:

part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB

Additional problem is the last part that is not labeled. Is it possible to do it with sed? Or maybe any other shell tools?

sed text bash regex
6 Answers

With GNU awk:

awk '{for(i=1; i<=NF; i++){print "part" i,"=",$i}}' FPAT='.{1,4}' file

Output:

part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB

From man awk:

FPAT: A regular expression describing the contents of the fields in a record. When set, gawk parses the input into fields, where the fields match the regular expression, instead of using the value of FS as the field separator.

user_37768580

The "hard" part of that, actually splitting the input into lines, is the job that fold exists to do:

$ echo 'AAAAAAAAABBBBBBBAAAAAAABBBBBBB' |
    fold -w 4 | awk '{print "part" NR " = " $0}'
part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB

Using fold will make your life easier if/when your requirements change to, say, try to only split at blanks.

user_17450010

Here is a simpler gnu awk solution:

s='AAAAAAAAABBBBBBBAAAAAAABBBBBBB'
awk -v RS='[^\n]{1,4}' 'RT{print "part" NR, "=", RT}' <<< "$s"

part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB
user_5482250

With GNU tools:

echo "$str" | grep -oE '.{,4}' | nl -s ' = ' | sed -E 's/^ +/part/'

part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB
user_75520

Another GNU awk, using builtin variable RT:

$ echo AAAAAAAAABBBBBBBAAAAAAABBBBBBB | 
awk -v RS=".{1,4}" '{printf "%spart%d = %s",(++i>1?ORS:""),i,RT}'

Sample output:

part1 = AAAA
part2 = AAAA
...
part8 = BB
user_41623560

In your code, the = in the replacement string of s command is just a literal = character, not a line number. This problem could be solved in pure sed using 3 sed commands in a pipeline like that:

line='AAAAAAAAABBBBBBBAAAAAAABBBBBBB'

echo "$line"                  |
sed 's/.\{4\}/&\n/g; s/\n$//' |
sed =                         |
sed 'N; s/\(.*\)\n\(.*\)/part\1 = \2/'

which outputs

part1 = AAAA
part2 = AAAA
part3 = ABBB
part4 = BBBB
part5 = AAAA
part6 = AAAB
part7 = BBBB
part8 = BB
user_138090010
Related