Anyone know the comand line to replace a string after a word for the 1st line?

i want to replace the number 266 with 100 for the first occurance after the date YYMMDD on the first line.

This is what i have so far

cDATE=`date '+%y%m%d'`

sed '/'${cDATE}'/,$ { 1s/266/100 }' <file> > <newfile>


however it keeps giving me garbled sed error and i can't for the life of me figure out what's wrong.

bump

I'm not an sed expert so I had to do some reading first. I'm not sure exactly what you want but I think this gets pretty close:

cDATE=`date '+%y%m%d'`
sed -e "/^$cDATE/,/^266/s/266/100/" file > file.new

This would substitute every occurance of 266 with 100 in a block starting with $cDATE and ending with 266.
But you said something about the first line. Maybe something like this would be more what you want:

sed -e "1s/\($cDATE\)\(.*\)266/\1\2100/" file > file.new

This would look for the pattern date-anything-266 and substitute it with firstmatchedpart-secondmatchedpart-100. Because of the 1 in the beginning it would only do it on the first line.

Oh yeah now I remember why that last one won't work. Sed matches the longest match so if you have 266 twice on that line only the last one will be replaced.

Thanks, that's exactly what i wanted. To match the first occurance on the first line after the 'date'.

Moocha gracia!

Noodle.

One last question, what if there's 2 occurances on that line? How would you modify u're 2nd example to change only the 1st occurnace?

I tried to figure that one out but couldn't really find a solution to that. The second example you can't modify to change only the first one. Because that's just the way sed works.
I managed to get the first example working only for the first line using a script file for the commands but then it won't substitute the $cDATE variable anymore. I'd have to read up on sed a bit more to figure that one out.

Thanks, u're example is working great for my script now, i just hope that there's just going to be one occurance on those first lines.

thanks.