How do I find the length of the longest line in a file?

[John Howard]

I'm not that great a perl programmer. Just know the basics.

I need some sample code to show me how to read a file and find the
length of the longest line in the file.

Thanks in advance.

 

[Juha Laiho]

I'm not that great a perl programmer. Just know the basics.

I need some sample code to show me how to read a file and find the
length of the longest line in the file.

How about thinking what needs to be done to achieve that - then
transform the thoughts into a program? I'll walk you through the
first part.

Very rough:
Looks like one variable is needed: the place to hold length of
longest line seen so far.

The file must be read a line at a time, comparing length of
each line to the length of longest line seen so far.

When all the lines have been read, the length of the longest
line is to be printed.

Adding some finer points to the above, and changing the wording
towards a computer program:
- declare a variable (let's call it $longest), initialize it to '-1'
- attempt to open the file, storing the error status (if any)
- if file open succeeded
- for the following lines ...
- read a line
- if length of line is greater than value stored in $longest,
store length of line to $longest
- ... repeat until end of file is reached

- if value of $longest is still '-1', output the error message
describing why the file opening failed
- otherwise, print a message telling that length of the longest
line in the file was $longest

.... and even now I feel I've been helping you all too much with
your homework.

 

[William James]

I'm not that great a perl programmer. Just know the basics.

I need some sample code to show me how to read a file and find the
length of the longest line in the file.

Thanks in advance.

awk 'length($0)>m{m=length($0)}END{print m}' somefile.txt

 

[James Taylor]

awk 'length($0)>m{m=length($0)}END{print m}' somefile.txt

Perl supports awkish syntax too:

perl -ne '$m=length if $m<length; END{print $m}' somefile.txt

 

[John W. Krahn]

I'm not that great a perl programmer. Just know the basics.

I need some sample code to show me how to read a file and find the
length of the longest line in the file.

my $file = 'somefile.txt';

open my $fh, '<', $file or die "Cannot open $file: $!";

my $longest;

while ( <$fh> ) {

$longest = length if $longest < length;
}

print "The longest line is $longest bytes.\n";




John

 

[William James]

Perl supports awkish syntax too:

perl -ne '$m=length if $m<length; END{print $m}' somefile.txt

Good point. Since "computer golf" is so much fun, here's my
solution in Ruby (I tried the awkish approach, but this is
shorter):

ruby -e 'p ARGF.map{|x|x.size}.max-1' somefile.txt

The -1 is there because of the trailing linefeed characters.

 

[Josef Moellers]

Very rough:
Looks like one variable is needed: the place to hold length of
longest line seen so far. [...]
- declare a variable (let's call it $longest), initialize it to '-1'

<Nitpicking mode>
IMHO, a "length" is _always_ positive, especially if it refers to
lengths of text-lines, so '-1' would be an illegal value.
I'd either leave it undef-ined (which would fit if you haven't seen any
lines yet) or set it to 0 (if you haven't seen anything, you have seen 0).
</Nitpicking mode>

 

[James Taylor]

Good point. Since "computer golf" is so much fun, here's my
solution in Ruby (I tried the awkish approach, but this is
shorter):

ruby -e 'p ARGF.map{|x|x.size}.max-1' somefile.txt

Is ARGF an array of lines of the files supplied on the command line
in a similar way to Perl's <> handle?

Doesn't Ruby's map method have the equivalent of Perl's anonymous
scalar? In Perl you might write:

@line_lengths = map length, <>;

without having to name a variable x *twice* as in your example.
So, just hand waving here, Ruby might write: AGRF.map{.size}
or similar.

PS. Is there any great motivation for your Ruby advocacy when we'll
all someday be sharing the Parrot runtime anyway? Seems to me the
main motivation for advocacy of one language over another would be
because you want more people to contribute to your favourite language's
codebase in the same way that Perl gains so much from CPAN contributions.
With Parrot, the language war is over because everyone's contributions
will work together regardless of the language they're written in.
So, I can't help feeling your efforts would be better directed in
the comp.lang.ruby group, rather than here. Ruby has too much in
common with Perl for it to be worthwhile any of us learning it if
we're already happy with Perl and, until Parrot is widely available,
CPAN tips the usefulness in favour of Perl anyway. Sure, you've shown
us that Ruby is cute, expressive, and powerful, but that's just
not enough for most people.

 

[Anno Siegel]

Very rough:
Looks like one variable is needed: the place to hold length of
longest line seen so far. [...]
- declare a variable (let's call it $longest), initialize it to '-1'

<Nitpicking mode>
IMHO, a "length" is _always_ positive, especially if it refers to
lengths of text-lines, so '-1' would be an illegal value.

Quite. Since -1 cannot be the length of any actual line, it serves
well to indicate the fact that no actual line has been seen yet.

I'd either leave it undef-ined (which would fit if you haven't seen any
lines yet) or set it to 0 (if you haven't seen anything, you have seen 0).
</Nitpicking mode>

Not so. You have seen 0 lines, but not a line of length 0, so you
have no reason to set the minimum length to any legal value, not even 0.

In practice, initializing the minimum to 0 would make the cases
indistinguishable where you have no lines at all and where you have
one or more lines all of length 0.

Using undef instead of -1 is of course okay (it is also not the length
of any line), but -1 makes the numeric treatment easier.

Anno

 

[axel]

Doesn't Ruby's map method have the equivalent of Perl's anonymous
scalar? In Perl you might write:

Who is Ruby and why should I wave to her... unless she
is a good-looking redhead which I doubt.

Axel

 

[John Howard]

my $file = 'somefile.txt';

open my $fh, '<', $file or die "Cannot open $file: $!";

my $longest;

while ( <$fh> ) {

$longest = length if $longest < length;
}

print "The longest line is $longest bytes.\n";

I ended up using a variant of this. Thanks. I didn't realise "length"
was the required function (the perl docs are rather hard to fathom. I
kept looking for something like strlen etc. Oh well).

Can it also be use for a field in a file? I.e. read the lines into
local vars using split and then check the length of one of those?

 

[Paul Lalli]

I ended up using a variant of this. Thanks. I didn't realise "length"
was the required function (the perl docs are rather hard to fathom. I
kept looking for something like strlen etc. Oh well).

They take some getting used to, I admit, but once you know how to work
them, I think it's generally pretty easy to find what you're looking
for. In this case:
perldoc perlfunc
would have given you a listing of all Perl functions, including a
section on Perl Functions by Category. The first category listed is
"Functions for SCALARs or strings", and length is near the beginning of
that list.

See also:
perldoc perltoc
for a full Table Of Contets of the Perl documentation.

Can it also be use for a field in a file? I.e. read the lines into
local vars using split and then check the length of one of those?

If I understand your question correctly, yes. length returns the
length in bytes of its argument. If no argument is provided, it
returns the length of the current value of $_.

#!/usr/bin/perl
use strict;
use warnings;
my $foo = "a bunch of word characters";
my @words = split / /, $foo;
for my $word (@words){
print "The length of '$word' is: ", length $word, "\n";
}
__END__

The length of 'a' is: 1
The length of 'bunch' is: 5
The length of 'of' is: 2
The length of 'word' is: 4
The length of 'characters' is: 10


Paul Lalli

 

[Jürgen Exner]

John Howard wrote:
[Solution snipped]

I ended up using a variant of this. Thanks. I didn't realise "length"
was the required function (the perl docs are rather hard to fathom. I
kept looking for something like strlen etc. Oh well).

Two notes:
- "How do I get the length of a string?" is the best known SAQ (Self
Answering Question) in the Perl community. There are many old threads about
it. Granted, it hasn't been asked for while now.
- You know about "perldoc perlfunc", section "Perl Functions by Category",
don't you?

Functions for SCALARs or strings
"chomp", "chop", "chr", "crypt", "hex", "index", "lc", "lcfirst",
"length", "oct", "ord", "pack", "q/STRING/", "qq/STRING/",
"reverse", "rindex", "sprintf", "substr", "tr///", "uc", "ucfirst",
"y///"

Can it also be use for a field in a file? I.e. read the lines into
local vars using split and then check the length of one of those?

In which way are "fields" different then "strings"?

jue

 

[Jürgen Exner]

If I understand your question correctly, yes. length returns the
length in bytes of its argument.

That would be a really bad bug. From "perldoc -f lenght":

length Returns the length in characters of the value of EXPR.

jue

 

[Paul Lalli]

That would be a really bad bug. From "perldoc -f lenght":

length Returns the length in characters of the value of EXPR.

Whoops. My mistake. Thanks for the correction, Jürgen.

Paul Lalli