if you need to know alignment length of the read on the reference after it aligned: Using bioperl modules

##mapped_read_Length.pl 
#!/usr/bin/perl -w

use strict;
use warnings;  
use Bio::DB::Sam;

my $usage = "Usage $0 <input.bam>\n";
my $infile = shift or die $usage;
my $sam = Bio::DB::Sam->new(-bam  => $infile);
my @chromosomes = $sam->seq_ids;   
print "Read_ID\tChr\tStart\tEnd\tMapped_Length\tCIGAR_String\n";

foreach my $chr (@chromosomes){
   my @alignments = $sam->get_features_by_location(-seq_id => $chr);
   for my $a (@alignments) {
      my $id  = $a->name;
      my $cigar  = $a->cigar_str;
      my $start  = $a->start;
      my $end    = $a->end;
      my $mapped_length = $end-$start;
     print "$id\t$chr\t$start\t$end\t$mapped_length\t$cigar\n";    
}
}
__END__

OUTPUT:

Read_ID    Chr    Start    End    Mapped_Length    CIGAR_String
BMW-S105R:17:H02AWADXX:2:1109:4198:22439    chr1    564455    564503    48    6M1I43M
BMW-S105R:17:H02AWADXX:2:1109:4309:28023    chr1    564682    564732    50    5M1D45M
BMW-S105R:17:H02AWADXX:2:1109:15415:28127    chr1    564689    564737    48    37M1I12M

There are many ways to skin a cat. Below is a c++ implementation of BamTools that will produce what your want. You can get much more information out the the al BamTools::Alignment object.

More about BamTools https://github.com/pezmaster31/bamtools

#include  "api/api_global.h"
#include  "api/BamReader.h"

using namespace std;
using namespace BamTools;


int main(int argc, const char * argv[])
{
   std::string filename = argv[1];
   BamTools::BamReader reader
   BamTools::BamAlignment al;
      while(reader.GetNextAlignment(al)){
            std::cout << al.Length << "\n";

        }
}

BEDOPS includes a bam2bed conversion script, which filters unmapped reads in the course of converting to BED. You can therefore get mapped read lengths by piping the sequence field of the converted data to awk, e.g.:

$ bam2bed < foo.bam \
    | cut -f 12 \
    | awk '{print length($0)}' \
    > mapped_read_lengths.txt

If you use samtools directly with alternatives suggested in other answers, you might first want to prefilter your data to mapped reads with GNU awk, before piping to downstream utilities:

$ samtools view foo.bam | gawk '(! and(4, $2))' > mapped_foo.sam
$ samtools view -S mapped_foo.sam | ...

Or, more simply:

$ samtools view foo.bam | gawk '(! and(4, $2))' | ...

Where ... denotes the use of downstream utilities.

If you are given bam files and want to know the READ length that was used in sequencing, this will work. replace bamfile by your bam file (and make sure it was indexed .. e.g. a bamfile.bam.bai exists)

Typically, I need to know this to run a pipeline.

All on one line.

samtools view bamfile.bam | head -n 1000 | gawk '{print length($10)}' | sort | uniq -c | perl -ane '$_ =~ s/^[ ]+//g;print $_' | sort -k 1nr,1nr | head -1 | cut -f2 -d " "