Very Computer

by **yan** » Fri, 18 Apr 2003 22:27:33

how to identify the first occurence of a patten by using sed or awk

e.g.: my file is as following. I want to identify
the lines which contains the first occurence of "\"

aa aaaaaaaaaa \
bbbbb cccc \
ddddd ddddddd
eeeeeeee eesss \
ffffff dddddddddd

so the line I want to be printed out are:
aa aaaaaaaaaa \
eeeeeeee eesss \

(note: "bbbbb cccc \" is not the one i want because the "\" is
the second
occurence, not the first occurence)

great thanks advance!

--
Posted via http://dbforums.com

by **j..** » Fri, 18 Apr 2003 23:52:53

> how to identify the first occurence of a patten by using sed or awk

> e.g.: my file is as following. I want to identify
> the lines which contains the first occurence of "\"

> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd
> eeeeeeee eesss \
> ffffff dddddddddd

> so the line I want to be printed out are:
> aa aaaaaaaaaa \
> eeeeeeee eesss \

You just need to keep a flag that tells when to print and when not
to. Here's an awk solution, but it can probably be done in the shell
too.

awk '
BEGIN {first = 1}
/\\/ {
if(first)
{
print $0
first=0
}
else
first = 1

Quote:}

' < data

by **Stephane CHAZELA** » Sat, 19 Apr 2003 00:39:21

> how to identify the first occurence of a patten by using sed or awk

> e.g.: my file is as following. I want to identify
> the lines which contains the first occurence of "\"

> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd
> eeeeeeee eesss \
> ffffff dddddddddd

> so the line I want to be printed out are:
> aa aaaaaaaaaa \
> eeeeeeee eesss \

[...]

sed -ne '/\\/!d;p;:1' -e '/\\/{n;b1' -e '}'

--
Stphane

by **yan** » Sat, 19 Apr 2003 03:23:06

sed -ne '/\\/!d;p;:1' -e '/\\/{n;b1' -e '}'

I got an error message. I thought may be typo but I check it very
carefully. thanks!

--
Posted via http://dbforums.com

by **yan** » Sat, 19 Apr 2003 03:19:09

sorry, the awk solution won't work.
if I have one more line as following:
aa aaaaaaaaaa \
bbbbb cccc \
ddddd ddddddd \
hhhhhhhh
eeeeeeee eesss \
ffffff dddddddddd

then it print out
aa aaaaaaaaaa \
ddddd ddddddd \

but i need
aa aaaaaaaaaa \
eeeeeeee eesss \

--
Posted via http://dbforums.com

by **j..** » Sat, 19 Apr 2003 04:17:02

> sorry, the awk solution won't work.
> if I have one more line as following:
> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd \
> hhhhhhhh
> eeeeeeee eesss \
> ffffff dddddddddd

> then it print out
> aa aaaaaaaaaa \
> ddddd ddddddd \

> but i need
> aa aaaaaaaaaa \
> eeeeeeee eesss \

Ok, I thought you wanted every other line containing '\'. If what you
really want is to print out lines containing '\' only after a line not
containing '\' adjust the counter accordingly.

awk '
BEGIN {first = 1}
{
if($0 ~ /\\/)
{
if(first)
{
print $0
first = 0
}
}
else
first = 1
}
' < data

by **William Par** » Sat, 19 Apr 2003 04:30:58

> how to identify the first occurence of a patten by using sed or awk

> e.g.: my file is as following. I want to identify
> the lines which contains the first occurence of "\"

> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd
> eeeeeeee eesss \
> ffffff dddddddddd

> so the line I want to be printed out are:
> aa aaaaaaaaaa \
> eeeeeeee eesss \

> (note: "bbbbb cccc \" is not the one i want because the "\" is
> the second
> occurence, not the first occurence)

> great thanks advance!

grep -A 1 | grep '\$'

man grep

--

Linux solution for data management and processing.

by **rakesh shar** » Sat, 19 Apr 2003 09:58:51

> how to identify the first occurence of a patten by using sed or awk

> e.g.: my file is as following. I want to identify
> the lines which contains the first occurence of "\"

> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd
> eeeeeeee eesss \
> ffffff dddddddddd

> so the line I want to be printed out are:
> aa aaaaaaaaaa \
> eeeeeeee eesss \

> (note: "bbbbb cccc \" is not the one i want because the "\" is
> the second
> occurence, not the first occurence)

> great thanks advance!

put these lines in a runfile, then invoke sed on them: sed -f runfile Input

:begin
/\\$/{
$b
:loop
N
/\\$/bloop
P
s/^.*\n//
bbegin

Quote:}

d

by **rakesh shar** » Sat, 19 Apr 2003 10:53:59

> how to identify the first occurence of a patten by using sed or awk

> e.g.: my file is as following. I want to identify
> the lines which contains the first occurence of "\"

> aa aaaaaaaaaa \
> bbbbb cccc \
> ddddd ddddddd
> eeeeeeee eesss \
> ffffff dddddddddd

> so the line I want to be printed out are:
> aa aaaaaaaaaa \
> eeeeeeee eesss \

> (note: "bbbbb cccc \" is not the one i want because the "\" is
> the second
> occurence, not the first occurence)

> great thanks advance!

there were some gotchas(eg,last 2 lines w/ \) in my previous version;

/\\$/!d
:loop
$!N
/\\$/{
$!bloop

Quote:}

P
d

by **John Savag** » Sun, 20 Apr 2003 08:33:36

>sed -ne '/\\/!d;p;:1' -e '/\\/{n;b1' -e '}'

>I got an error message. I thought may be typo but I check it very
>carefully. thanks!

And that error message was exactly what?

Some seds require labels to stand alone, so try rearranging as:

sed -ne '/\\/!d;p' -e ':1' -e '/\\/{n;b1;}'
--
John Savage (newsgroup email invalid; keep news replies in newsgroup)

Very Computer

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help

sed, awk help