Search Results

$ awk '!(SEEN[$1,$4]++) && !(SEEN[$4,$1])' <<EOF
dmn10003t1 PF00001 PF00022 dmn12390t1
dmn10008t1 PF00069 PF00027 dmn9781t1
dmn10008t1 PF00068 PF00027 dmn9781t1
dmn10008t1 PF00069 PF00069...

How about awk ' {c++
for (i=1; i<=NF; i++) if ($i ~ /prt1/) A[$i]
else B[$i]
for (i in A) for (j in...

That space between block 2 and 3 hoses zaxxon's fine proposal... plus - I'm not sure what the FS is for.

$ awk '{print $0 > "cluster"NR".txt"}' FS="\n+" RS= infile

The OP did not specify an operating system, so I'll mention that because of \n in the replacement text, that sed script will fail on most non-Linux systems.

Regards,
Alister

---------- Post...

I do see now that the OPs output request also has space in front
awk '{$1=" "$1}1' OFS="\n " ORS="\n\n"
This gives space in front of all line.
But we do not know for sure of is this he wants :)

sed has same problem as perl
it adds double blank line, since there is a space infront of the line
abc
qwe
tyu
ghj
jkl
dfg
sdf


cvb
sdk
fgh
tyu


uio
...
...

sed works too ...
sed -e "s/$/\n/" -e "s/[ \n]/\n/g" infile

awk '{$1=$1}1' OFS="\n" ORS="\n\n" infile
This awk works if you have a space in front of some of the line or not
cvb vs cvb
@durden_tylor, you have removed this space in f02
perl gives an extra...

$
$ cat f02
abc qwe tyu ghj jkl dfg sdf
cvb sdk fgh tyu
uio iop tyu rty eru wer rty iop
asd sdf dfg fgh zxc
$
$ perl -pne 'BEGIN {$\="\n"}s/ /\n/g' f02
abc
qwe
tyu
ghj
jkl
dfg
sdf
...

sed '
s/\( *[^ ]\{1,\}\)/\1\
/g
' infile