inplace editing (-i) in Sed

[Azrael2]

Sed is like really old, but for some reason, the sed version on one of the boxes I work with, doesn't support inplace editing. this is annoying because you then generate a new file every time you run it through sed.

I guess you can't specify more than one "find and replace" per sed command, can you? Or there's no trick to it?

I'm not sure I should go and upgrade the version of sed on there in case it breaks something... any ideas?

*edit* I just found replace, but as I want to replace a load of ^Ms with nothing, it won't work :/

[anomie]

the sed version on one of the boxes I work with

What OS / version?

I guess you can't specify more than one "find and replace" per sed command, can you?

Yes you can; see the -e option.

I want to replace a load of ^Ms with nothing

One option:
awk '{ sub("\r$", ""); print; }' your_file_here > new_file_here

Another option:
dos2unix your_file_here

Another option:
Search the forums. This has been discussed before.

[Azrael2]

The best way I found is perl -pi -e and then the same syntax you would use for Sed. Works a treat, and is much tidier. thanks anyway

[kakariko81280]

There is a program called sponge, which lets you do things like:

sed 'pattern' /file | sponge /file

without clobbering /file.

Chris...