Substituting the first occurrence of a pattern in a line, for all the lines in a file with sed

regular expressionsed

Is it possible to do it in one liner?

I have an output like this:

"First line" - Description  
" Second line" - Description  
"Third line" - Description  
" Fourth line" - Description

This input is generated automatically.

I want to replace the first occurrence of " (quotation mark + space) with " (quotation mark) for each line. If I apply the substitution globally, it will also change every occurrence of line" - to line"-, so I was wondering if it is possible to use a sed one liner to accomplish this.

I have tried using ^ like this

sed -r ':a;N;$!ba;s/(\^\" )/\"/g'

But it's not working, it doesn't replace anything. I tried

sed -r ':a;N;$!ba;s/(^|\" )/\"/g'

and it replaces all the occurrences. I've just started to use sed, so I don't really know if I'm doing something wrong.

What am I missing here?

Best Answer

You're overthinking it. sed replaces only the first instance on a line by default (without the /g modifier), although you still want to anchor because you don;t so much want the first instance in the line as the one at the start of the line; and you usually don't need the explicit line actions you're trying to use (why?).

sed 's/^" /"/'

Related Solutions

Using sed to find and replace complex string (preferrably with regex)

sed -i -E "s/(<username>.+)name(.+<\/username>)/\1something\2/" file.xml

This is, I think, what you're looking for.

Explanation:

parentheses in the first part define groups (strings in fact) that can be reused in the second part
\1, \2, etc. in the second part are references to the i-th group captured in the first part (the numbering starts with 1)
-E enables extended regular expressions (needed for + and grouping).

Search and replace using grep (not sed)

grep is only meant to (and was only initially) print(ing) the lines matching a pattern. That's what grep means (based on the g/re/p ed command).

Now, some grep implementations have added a few features that encroaches a bit on the role of other commands. For instance, some have some -r/--include/--exclude to perform part of find's job.

GNU grep added a -o option that makes it perform parts of sed's job as it makes it edit the lines being matched.

pcregrep extended it with -o1, -o2... to print what was matched by capture groups. So with that implementation, even though it was not designed for that, you can actually replace:

sed 's/old/new/'

with:

pcregrep --om-separator=new  -o1 -o2 '(.*?)old(.*)'

That doesn't work properly however if the capture groups match the empty string. On an input like:

XoldY
Xold
oldY

it gives:

XnewY
X
Y

You could work around that using even nastier tricks like:

PCREGREP_COLOR= pcregrep --color=always '.*old.*' |
  pcregrep --om-separator=new -o1 -o2 '^..(.+?)old(.+)..' |
  pcregrep -o1 '.(.*).'

That is, prepend and append \e[m (coloring escape sequence) to all matching lines to be sure there is at least one character on either side of old, and strip them afterwards.

Best Answer

Related Solutions

Using sed to find and replace complex string (preferrably with regex)

Search and replace using grep (not sed)

Related Question