Very Computer

Hi,

I want to search for expressions in a file which include a new-line. Is
this possible with grep or egrep. Or is there another solution with
shell utilities.

thanks in advance

Heiko

Quote:>Hi,

>I want to search for expressions in a file which include a new-line. Is
>this possible with grep or egrep. Or is there another solution with
>shell utilities.

As usual, with Unix, there are several ways in which you can do this.
If there is a character in your file which occurs with high frequency,
but is not part of your pattern, (space is a good candidate for text files)
then you can use 'tr' to swap between your character and newline, e.g.

e.g. if you are looking for pat\nern

tr ' \n' '\n ' < file | grep 'pat ern'

This is mostly useful if you are just trying to establish the existance of
the match, rather than seeing the rest of the lines.

You can use awk

awk '/^ern/ { if (seenpat) { print saved; print $0 } }
        { if ($0 ~ /pat$/) { saved=$0 ; seenpat=1 } else { seenpat=0 } }'

This works by using a variable seenpat to remember if pat ended the previous
line, and printing the line only if the new line starts with ern and the
previous line ended with pat.

If you have perl installed, you can set it to read in a paragraph at a time
and match over newlines

perl -00ne 'print if (/pat\nern/);'

or you could use sed.
sed -e 'N;/pat\nern/p;D'

This works as follows
The 'N' joins the next line of input onto the 'pattern space'. It includes
a newline as a seperator.
If we see the pattern, then we print out the pattern space.
The 'D' deletes everything up to the first newline from the pattern space,
and causes us to loop back to the beginning.

If you have 'agrep' (approximate grep) installed, then you can set it to
paragraph mode.

I would use perl, because it comes most naturally to mind.
Icarus