SED replace few first occurences ( and ranges ) of pattern

Question

is this possible to change first 4 ( or more ) occurences of string in this scenario using SED (opposite of sed -r 's/[^[:space:]]*/TEST/4g'):

TEST TEST TEST TEST five six seven

I get it working with reversing words order in line using AWK twice, but this is long, complex and I want to get this with just SED:

echo one two three four five six seven | awk '{for(i=NF;i>=1;i--) printf "%s ", $i;print ""}'  | sed -r 's/[^ ]*/TEST/4g' |  awk '{for(i=NF;i>=1;i--) printf "%s ", $i;print ""}'

Also maybe there is option to change ranges of occurence like 3-5, 6-12, ...?

Example input is:

one two three four five six seven

eight nine ten eleven twelve thirteen fourteen

fifteen sixteen seventeen eighteen nineteen twenty twenty-one

Answer 1

What about a single AWK:

awk '{for(i=1;i<=NF;i++) if(i<5){$i="TEST"}; print}'

Test run:

$ echo one two three four five six seven | awk '{for(i=1;i<=NF;i++) if(i<5){$i="TEST"}; print}'
TEST TEST TEST TEST five six seven

This solution is short, readable and maintainable. If it does not satisfy you, please add some details about your specific problem.

---

Perl equivalent solution:

perl -pe 's/\S+/$i++<4?"TEST":$&/ge'

Test run:

$ echo one two three four five six seven | perl -pe 's/\S+/$i++<4?"TEST":$&/ge'
TEST TEST TEST TEST five six seven

---

maybe there is option to change ranges of occurence like 3-5, 6-12

AWK:

awk '{for(i=3;i<6;i++)$i="TEST";print}'

Test run on the newly provided input file:

$ awk '{for(i=3;i<6;i++)$i="TEST";print}' input
one two TEST TEST TEST six seven
eight nine TEST TEST TEST thirteen fourteen
fifteen sixteen TEST TEST TEST twenty twenty-one

Perl:

perl -pe 's/\S+/++$c~~[3..5]?"TEST":$&/ge'

Test run on the newly provided input file:

$ perl -pe '$c=0;s/\S+/++$c~~[3..5]?"TEST":$&/ge' input
Smartmatch is experimental at -e line 1. <== This is a warning that goes to STDERR
one two TEST TEST TEST six seven
eight nine TEST TEST TEST thirteen fourteen
fifteen sixteen TEST TEST TEST twenty twenty-one

Answer 2

The answer has been provided here by mikeserv. NOTE: if you want to process a range, you need to use the maximum bound, as it will process as many matches as it can without throwing any exceptions/errors.

GNU sed:

echo 'one two three four five six seven' | \
  sed 's/[^[:space:]]*/\n&/g;:t;/\n/{x;/.\{4\}/!{s/$/./;x;s/\n[^[:space:]]*/TEST/;bt};x};s/\n//g'

POSIX sed:

nl='
';
echo 'one two three four five six seven' | sed "s/[^[:space:]]*/\\$nl&/g;:t${nl}/\n/{x;/.\{4\}/!{${nl}s/$/./;x;s/\n[^[:space:]]*/TEST/;bt$nl};x$nl};s/\n//g"

See the online sed demo.

Original explanation (note that here, 1 is replaced with 2, you may use any other patterns):

There I use two notable techniques. In the first place every occurrence of 1 on a line is replaced with \n1. In this way, as I do the recursive replacements next, I can be sure not to replace the occurrence twice if my replacement string contains my replace string. For example, if I replace he with hey it will still work.

I do this like:

s/1/\
&/g

Secondly, I am counting the replacements by adding a character to hold space for each occurrence. Once I reach three no more occur. If you apply this to your data and change the \{3\} to the total replacements you desire and the /\n1/ addresses to whatever you mean to replace, you should replace only as many as you wish.

Answer 3

This is a completely inappropriate task for sed as sed is for doing simple s/old/new/ on individual strings, that is all. With any awk in any shell on every UNIX box:

$ echo one two three four five six seven | awk '{for (i=1; i<=4; i++) $i="TEST"}1'
TEST TEST TEST TEST five six seven

$ echo one two three four five six seven | awk '{for (i=3; i<=5; i++) $i="TEST"}1'
one two TEST TEST TEST six seven

and if you need to parameterize it:

echo one two three four five six seven |
    awk -v beg=3 -v end=5 '{for (i=beg; i<=end; i++) $i="TEST"}1'
one two TEST TEST TEST six seven

Answer 4

$ echo "one two three four fix six" | \
sed -E ':r s/(^|(TEST )+)[^ ]*/\1TEST/;/^(TEST ){4}/!br'
TEST TEST TEST TEST fix six

Explanation:

• :r label named r to branch back to
• s/(^|(TEST )+)[^ ]*/\1TEST/; replacement that replaces just one occurrence of a non-TEST word, preceeded by either the start of the line or 1 or more TESTs
• /^(TEST ){4}/!br' regex for what's wanted, followed by the !br to branch back to :r if it's not matched yet.

Clearly this is fragile. It will loop infinitely if any lines don't have four words. Might be GNU sed only.