How can I use awk to insert something in the middle of the word?

Question

I have an input:

This is a test

And I want to insert some letters in the middle of the word, like:

This is a teSOMETHINGst

I know I can define the needed word by $i, but how can I modify the word that way?

I'm trying to do it like that:

{
    i=4 # finding somehow
    print (substr($i,1,(length($i)/2)) "SOMETHING" substr($i,(length($i)/2),(length($i)/2)))
}

As I'm new to awk I wonder if it is a right way.

Answer 1

This may be what you're looking for:

$ awk 'match($0,/\<test\>/){mid=int(RLENGTH/2); $0=substr($0,RSTART,mid) "SOMETHING" substr($0,RSTART+mid,RELNGTH-mid)} 1'

e.g. some test cases (no pun intended):

$ echo 'This is a test' |
awk 'match($0,/\<test\>/){mid=int(RLENGTH/2); $0=substr($0,RSTART,mid) "SOMETHING" substr($0,RSTART+mid,RLENGTH-mid)} 1'
teSOMETHINGst

$ echo 'These are tests' |
awk 'match($0,/\<tests\>/){mid=int(RLENGTH/2); $0=substr($0,RSTART,mid) "SOMETHING" substr($0,RSTART+mid,RLENGTH-mid)} 1'
teSOMETHINGsts

$ echo 'These contestants are in a test' |
awk 'match($0,/\<test\>/){mid=int(RLENGTH/2); $0=substr($0,RSTART,mid) "SOMETHING" substr($0,RSTART+mid,RLENGTH-mid)} 1'
teSOMETHINGst

Answer 2

Assuming your requirement is to match the column number containing test and do some operations over it, do a simple loop over the columns upto NF and match using the regex match operator ~ or for fixed strings do a equality match as $i == "test"

awk '
{
  for(i=1;i<=NF;i++) {
    if ($i ~ "test") {
      halfLength=(length($i)/2)
      $i=(substr($i,1,halfLength) "SOMETHING" substr($i,(halfLength+1),halfLength))
    }
  }
}1' <<<"This is a test"

This produces the output as expected. Note that I've made the substr() call for printing the 2nd part of the string as substr($i,(halfLength+1),halfLength). The +1 is needed which you have missed before. I've used the substr() result to be modify column number containing test i.e. as $i=..

Also when doing {..}1, each of the column fields are reconstructed based on the modifications if any, in our case only to the column containing the string you wanted.

Also note that the whole attempt will fail if the target string contains an odd number of characters or forms a sub string of another larger string ( could use the equality operator but regex approach would fail )

Answer 3

Another another one that grew from curiosity to personal vendetta (:

$ echo This is a contestant test | 
awk -v s="test" '
BEGIN {
    FS=OFS=""
}
{
    if(i=match($0, "(^| )" s "( |$)")) {   # match over index since regex support
        j=(i+length(s)/2+!!(i-1))          # !!(i-1) detect beginning of record
        $j="SOMETHING" $j
    }
}1'
This is a contestant teSOMETHINGst

Another one using empty separators, mostly to satisfy personal curiosity:

$ echo This is a test | 
awk -v s="test" '
BEGIN {
    FS=OFS=""                # empty separators
}
{
    if(i=index($0,s)) {      # index finds the beginning of test
        j=(i+length(s)/2)    # midpoint
        $j="SOMETHING" $j    # insert string
    }
}1'                          # output
This is a teSOMETHINGst