is there a way to replace a new line with sed?-CodePudding

I have this text.

Code=200]:[ (381_3544|null|https_//cdn10.connected-stories.com)(
             Penn State World Campus  - San Deigo
        |TV)]:677:01:SUCCESS
20220929010707:[url=https_//rtr.innovid.com/r1.62b49edb185030.50315160;cb=1372882909, breakId=1

I want to remove a new line \n end of "(" and "Deigo" Example:

Code=200]:[ (381_3544|null|https_//cdn10.connected-stories.com)(Penn State World Campus  - San Deigo|TV)]:677:01:SUCCESS
20220929010707:[url=https_//rtr.innovid.com/r1.62b49edb185030.50315160;cb=1372882909, breakId=1

I tried using this command.

sed -ie "s/San Deigo\n/San Deigo/g"

It is not working. I also tried -z option but it deleted all the line and append to the first line.

CodePudding user response：

Using GNU sed

$ sed -E ':a;N;s/((\()|(Deigo))\n  /\2\3/;ba' input_file
Code=200]:[ (381_3544|null|https_//cdn10.connected-stories.com)(Penn State World Campus  - San Deigo|TV)]:677:01:SUCCESS
20220929010707:[url=https_//rtr.innovid.com/r1.62b49edb185030.50315160;cb=1372882909, breakId=1

CodePudding user response：

With your shown samples please try following GNU awk. Written and tested with shown samples only.

awk -v RS='\n[0-9]{14}:\\[url=https?_' '
FNR==1{
  gsub(/\n/,"")
  printf("%s%s",$0,RT)
}
END{
  printf("%s",$0,"")
}
'  Input_file

CodePudding user response：

sed by default processes one line at a time; but you can join more than one line into the pattern space and then replace it.

sed -e '/San Deigo$/N;s/\n//' file

If you want to learn sed, probably spend half an hour reading an introduction to its facilities. The Stack Overflow sed tag info page has links to resources.

Generally, though, perhaps you are better off using a language which is not write-only.

awk '/San Deigo$/ { printf "%s", $0; next }1' file

Awk out of the box does not have the -i flag which is supported by many sed variants (though it's not standard POSIX); GNU Awk has a -i inplace plug-in but then you need to make sure your Awk is gawk.

CodePudding user response：

With perl:

perl -pe 'chomp if /\($|San Deigo$/'

This will only remove the newline character. To also remove starting whitespace of lines being joined:

perl -pe 's/^\s // if $c==$.; if(/\($|San Deigo$/){$c=$. 1; chomp}'

Use perl -i for inplace editing.

Also, don't use sed -ie, it will create a backup file of the original with e appended to the name (unless that's what you wanted).

CodePudding user response：

This might work for you (GNU sed):

sed -E ':a;N;s/((\(|Deigo))\n/\1/;ta;P;D' file

Append the next line to the pattern space.

If the first line in the pattern space ends in ( or Deigo followed by a newline, remove the new line and go again by jumping to the label :a.

Othewise, print/delete the first line in the pattern space and repeat.

N.B. The D command deletes upto and including the first newline or if there is no newline, deletes the entire line. Unlike the d command it alters the sed cycle in that if there had been a newline it does not read the next line but begins the cycle as if it had and continues with the remnants of the pattern space from the previous cycle.