sed to delete a line with fixed start

[Bill McNAMARA_1]

Hi,

Got a lot of xml files that I need to delete a line out of before my browser opens them.
The line always starts with
but the rest will always be different.

How can I use sed to do a delete or ' ' replacement of the entire line based on this.

Thanks guys!
Bill

[John Poff · Accepted solution]

Hi Bill,

Try this:

sed '/^


JP

[John Poff]

Hi,

That last example works for being at the beginning of the line. If you want to delete the lines that contain anywhere on the line, just try it without the caret symbol:

sed '/


JP

[Joseph C. Denman]

Bill,

I think this will work.

sed '/^ filename.out
mv filename.out filename

...jcd...

[Thierry Poels_1]

hi,
how about "grep -v ..." ?
regards,
Thierry.

[Bill McNAMARA_1]

Thanks guys,
it was just a simple cut and paste,
you saved me a lot of time... (I can go home
early now!)
Thanks,
have a good weekend.

Later,
Bill

[Thierry Poels_1]

grep -v '^ $F ;-)
regards,
Thierry.

[Joseph C. Denman]

Bill,

I really don't know what is happening. I cut and paste your script and your file. Ran your script and everything worked fine as expected??????????

If adding a line to the end fixes your problem, then add it. Before the sed's in your script do:

echo "" >> $F


...jcd...

[Bill McNAMARA_1]

yea, I'm baffled too.

after further analysis it appears to be only happening to files with encoding="UTF-8"
possibly related to the...
sed "s/$OLD/$NEW/g" $F > ${F}.tmp
I'll remove the rm just to verify and get back to you all.

Later,
Bill

[Bob_Vance]

Yeah. Could be construed as a bug, I suppose.
'sed' will drop the last "line" of a file if it has not been terminated by a newline char (I guess that it's not really a "line", then, is it :).
'grep -v', as was suggested, doesn't suffer from this malady, however.