Prev: how to log comman to a file ?
Next: Question using sed replace text file \t
From: Rakesh Sharma on 22 Apr 2010 10:58
On Apr 22, 3:38 pm, ezhil <ezhi...(a)gmail.com> wrote:
> Hi,
>
> I have a file with 12 fields and 3000 row. For each row, I would like
> to print col 1 and then when it comes to col 7, it should print 2
> columns together followed by a field separation till the end. I have
> tried something like:
>
> awk '{printf("%s\t", $1); for(i=7; i<NF; (i+2)) {printf("%s%s\t",$i, $
> (i+1) ) }; {printf("\n")} }' file1
>
> This prints only the col 1 and col 7 & 8 for the first record (many
> times and I have to use ctrl-c to interrupt). Could you please help me
> to fix this?
>
> Thanks in advance.
>
> Kind regards,
> Ezhil


Assumptions for 'sed':
i) only spaces no TABS as whitespace.
ii) no leading & trailing spaces.
iii) no multiple spaces

sed -e '
h ;# store the line for reuse

;# extract Col1 & store it for later use
s/ /\
/;s/\n.*//;x

;# strip Cols1-6
s/ /\
/6;s/.*\n//

;# label odd space occurence with newline & even with tab
:loop
s/ /\n/
s//\t/
tloop
s/\n/ /g ;# then change all newlines back to space

H;g;s/\n/\t/ ;# join col1 to cols7,8, onwards.
' yourfile

perl -pale '
$_ = $F[0];
shift @F for 1..6;
$_ = join "\t",$_,map{"$F[2*$_] $F[2*$_+1]"} 0..$#F/2;
' yourfile

### note: assuming an even number of columns.

untested.

--Rakesh
From: Janis Papanagnou on 23 Apr 2010 06:37
Rakesh Sharma schrieb:
> On Apr 22, 3:38 pm, ezhil <ezhi...(a)gmail.com> wrote:
>> Hi,
>>
>> I have a file with 12 fields and 3000 row. For each row, I would like
>> to print col 1 and then when it comes to col 7, it should print 2
>> columns together followed by a field separation till the end. I have
>> tried something like:
>>
>> awk '{printf("%s\t", $1); for(i=7; i<NF; (i+2)) {printf("%s%s\t",$i, $
>> (i+1) ) }; {printf("\n")} }' file1
>>
>> This prints only the col 1 and col 7 & 8 for the first record (many
>> times and I have to use ctrl-c to interrupt). Could you please help me
>> to fix this?
>>
>> Thanks in advance.
>>
>> Kind regards,
>> Ezhil
>
>
> Assumptions for 'sed':
> i) only spaces no TABS as whitespace.
> ii) no leading & trailing spaces.
> iii) no multiple spaces

Why a "solution" with such strong restrictions if the OP's original
awk approach (modulo the simple errors in his code) doesn't require
any such "assumptions".

>
> sed -e '
> h ;# store the line for reuse
>
> ;# extract Col1 & store it for later use
> s/ /\
> /;s/\n.*//;x
>
> ;# strip Cols1-6
> s/ /\
> /6;s/.*\n//
>
> ;# label odd space occurence with newline & even with tab
> :loop
> s/ /\n/
> s//\t/
> tloop
> s/\n/ /g ;# then change all newlines back to space
>
> H;g;s/\n/\t/ ;# join col1 to cols7,8, onwards.
> ' yourfile

This is not yet cryptic enough; any chance to provide an Intercal
solution?

>
> perl -pale '
> $_ = $F[0];
> shift @F for 1..6;
> $_ = join "\t",$_,map{"$F[2*$_] $F[2*$_+1]"} 0..$#F/2;
> ' yourfile
>
> ### note: assuming an even number of columns.
>
> untested.
>
> --Rakesh
 | 
Pages: 1
Prev: how to log comman to a file ?
Next: Question using sed replace text file \t