What Pierre said. Assuming you have dumped to fasta file myseqs.fa, here's a Bioperl approach:

#!/usr/bin/perl -w

use strict;
use Bio::SeqIO;

my $inseq  = Bio::SeqIO->new(-file => "myseqs.fa", -format => "fasta");
my $outseq = Bio::SeqIO->new(-file => ">myseqs200.fa", -format => "fasta");

# write to file myseqs200.fa
while(my $seq = $inseq->next_seq) {
  if($seq->length <= 200) {
    $outseq->write_seq($seq);
  }
}

Use fastacmd to dump your database as fasta (option -D 1 ) and filter the result using your favorite tool (perl, awk, etc... )