Re: Manipulating file content
On Wed, Jun 28, 2000 at 02:37:33PM -0500, Bolan Meek wrote:
Viktor Rosenfeld wrote:
Hi list,
this one is for all the regexp, shell, and editing-experts...
How about us perl hackers, hunh?! Got sumpin' g'inst us, buddy!?
seems to me, us perl folk are the culmination of the cross-breeding
of all those geni. not that that's a good thing... :)
Suppose I have a comma-seperated or tab-seperated file and I want
to flip the lines and columns. So an input file like:
a,1,A
b,2,B
c,3,C
would be transformed into:
a,b,c
1,2,3
A,B,C
Is there an fast and easy way of doing this? Ideally through the shell
or with VIM. I would RTFM, but I have no clue which manual to read.
this is the kind of thing perl was designed for.
[OT in a large way]
i'd try something like (tmtowtdi):
#!/usr/bin/perl
$sep = shift || \t; # can supply separator on cmd line
@ary = (); # for peace of mind
for ($ln=0; ; $ln++) {
chomp;
$fn = 0;
foreach (split(/$sep/o)) {
$ary[ $fn++ ][ $ln ] = $_;
#$ary[col][row] == val
}
}
print join(\n, map {join($sep,@$_)} @ary ),\n;
alarmingly, it even seems to work well on mismatched tables
such as
a b c
1 2 3 4 5
i ii
A B C D
Re: Manipulating file content
Hello Bolan, Bolan Meek wrote: Here's the replacement code, in a script, instead of a one-liner. (I could've one-lined it, but this'll be more understandable: look! I even added _comments_) [Perl code snipped] This works perfect, well, almost perfect. It inserts a space after each comma. The fix is a no-brainer of course ($newRow[$i] .= $entry[$i],; in the first loop and a $newRow[$j] .= at the beginning of the second). Thanks heaps, Viktor -- Viktor Rosenfeld E-Mail: mailto:[EMAIL PROTECTED] HertzSCHLAG:http://www.informatik.hu-berlin.de/~rosenfel/hs/
Re: Manipulating file content
Will Trillich wrote: i'd try something like (tmtowtdi): [Perl code snipped] This one worked perfectly right out of the box. With the added value that I can supply a seperator on the command line :) and the disadvantage that I have no clue, what the code is actually doing :( [1] Big thank you, Viktor [1] It's not deleting my files, though. That's good. :) -- Viktor Rosenfeld E-Mail: mailto:[EMAIL PROTECTED] HertzSCHLAG:http://www.informatik.hu-berlin.de/~rosenfel/hs/
Re: Manipulating file content
Bolan Meek wrote:
this one is for all the regexp, shell, and editing-experts...
How about us perl hackers, hunh?! Got sumpin' g'inst us, buddy!?
Of course not! How could I?! :)
Well, you could use regexp in sed, or use an awk script, but if
I had only 3x3 matrices to transform, in text, I'd
perl -e 'for ($i=0;$i3;++$i){;@entry = split ',';print
$i[0],$i[1],$i[2]\n;}'
with a file directed into it, and stdout redirected to a file.
This ..., well ..., it doesn't work. At first I thought that you meant
@entry[x] in your last line, but that doesn't help either. I always get
3 pairs of commas without the values. Besides, the way I read the code,
it doesn't do anything usefull, because a line with values seperated by
commas, will become exactly the same line. But then again, I don't know
anything about Perl, so this is just guessing.
Matrixes with unpredetermined columns or rows become slightly
trickier, but only by 1) keeping track of the length/breadth, and
2) nesting another loop.
What about matrixes with a different number of columns and rows (e.g.
4x3 or 123x234)?
MfG Viktor
--
Viktor Rosenfeld
E-Mail: mailto:[EMAIL PROTECTED]
HertzSCHLAG:http://www.informatik.hu-berlin.de/~rosenfel/hs/
Re: Manipulating file content
Viktor Rosenfeld wrote:
Bolan Meek wrote:
this one is for all the regexp, shell, and editing-experts...
How about us perl hackers, hunh?! Got sumpin' g'inst us, buddy!?
Of course not! How could I?! :)
Well, we members of the Perl Hackers Anti-Defamation League
sometimes are a little touchy...
Well, you could use regexp in sed, or use an awk script, but if
I had only 3x3 matrices to transform, in text, I'd
perl -e 'for ($i=0;$i3;++$i){;@entry = split ',';print
$i[0],$i[1],$i[2]\n;}'
with a file directed into it, and stdout redirected to a file.
This ..., well ..., it doesn't work.
(Head under the arm) Guess I ought have _tested_ that first, hunh?
At first I thought that you meant @entry[x] in your last line,
Actually, I meant $entry[x]...
but that doesn't help either.
Yes, I see that now.
I always get 3 pairs of commas without the values.
Besides, the way I read the code,
it doesn't do anything usefull, because a line with values seperated by
commas, will become exactly the same line. But then again, I don't know
anything about Perl, so this is just guessing.
No, you're right. That was a quick stupid of me.
Matrixes with unpredetermined columns or rows become slightly
trickier, but only by 1) keeping track of the length/breadth, and
2) nesting another loop.
What about matrixes with a different number of columns and rows (e.g.
4x3 or 123x234)?
OK. I'll figure out why I'm not getting from my split what I
expect, correct my script, extend it for arbitrary matrices,
and get back to you.
(Boy, do I feel foolish...) That's what I get for not actually
_testing_ my code. I guess I'd better resign from the PHADL,
since I'm going to give us a bad name
Re: Manipulating file content
Viktor Rosenfeld wrote:
Bolan Meek wrote:
this one is for all the regexp, shell, and editing-experts...
How about us perl hackers, hunh?! Got sumpin' g'inst us, buddy!?
Maybe I ought have said perl slackers. (Well, I can get
away with that since I resigned from PHADL).
Well, you could use regexp in sed, or use an awk script, but if
I had only 3x3 matrices to transform, in text, I'd
WARNING! BAD CODE! WARNING!
perl -e 'for ($i=0;$i3;++$i){;@entry = split ',';print
$i[0],$i[1],$i[2]\n;}'
with a file directed into it, and stdout redirected to a file.
Here's the replacement code, in a script, instead of a one-liner.
(I could've one-lined it, but this'll be more understandable:
look! I even added _comments_)
#!/usr/bin/perl
while ()
{
chop; #chop \n
$origCols = (@entry = split /,/);
for ($i = 0; $i $origCols; ++$i)
{
$newRow[$i] .= $entry[$i], ;
}
}
for ($j = 0; $j $i; ++$j)
{
chop $newRow[$j]; # chop trailing ' '
chop $newRow[$j]; # chop trailing ','
print $newRow[$j]\n;
}
What about matrixes with a different number of columns and rows (e.g.
4x3 or 123x234)?
I tested it with a 5x3 matrix. It ought to handle arbitrary sizes,
but it doesn't test for irregular matrices, you know: something like
1,2,3,4,5
A,B,C,D,E,F,G
a,b,c
I,II,III,IV,V
Manipulating file content
Hi list, this one is for all the regexp, shell, and editing-experts on this list. Suppose I have a comma-seperated or tab-seperated file and I want to flip the lines and columns. So an input file like: a,1,A b,2,B c,3,C would be transformed into: a,b,c 1,2,3 A,B,C Is there an fast and easy way of doing this? Ideally through the shell or with VIM. I would RTFM, but I have no clue which manual to read. TIA, Viktor -- Viktor Rosenfeld E-Mail: mailto:[EMAIL PROTECTED] or: mailto:[EMAIL PROTECTED] HertzSCHLAG:http://www.informatik.hu-berlin.de/~rosenfel/hs/
Re: Manipulating file content
Viktor Rosenfeld wrote:
Hi list,
this one is for all the regexp, shell, and editing-experts...
How about us perl hackers, hunh?! Got sumpin' g'inst us, buddy!?
Suppose I have a comma-seperated or tab-seperated file and I want
to flip the lines and columns. So an input file like:
a,1,A
b,2,B
c,3,C
would be transformed into:
a,b,c
1,2,3
A,B,C
Is there an fast and easy way of doing this? Ideally through the shell
or with VIM. I would RTFM, but I have no clue which manual to read.
Well, you could use regexp in sed, or use an awk script, but if
I had only 3x3 matrices to transform, in text, I'd
perl -e 'for ($i=0;$i3;++$i){;@entry = split ',';print
$i[0],$i[1],$i[2]\n;}'
with a file directed into it, and stdout redirected to a file.
Matrixes with unpredetermined columns or rows become slightly
trickier, but only by 1) keeping track of the length/breadth, and
2) nesting another loop.
TIA,
YWATF
(You're Welcome After The Fact)
