PLINK 1.9 has --recode oxford for direct export, and --data/--gen/--bgen/--sample for import. --hard-call-threshold can be used to set a genotype likelihood cutoff, or randomize genotypes based on the likelihoods, during import.

I have had this problem... I wrote a quick and dirty perl script to get past this. Nothing fancy, but it works. This is meant to do this by chromsome - I split these up on a cluster computer, but hopefully this gets you going.

#!/usr/bin/perl -w

$file = $ARGV[0];
$chr = $ARGV[1];

open(FILE,"<$file") || die;

while(<FILE>) {
    chomp($_);
    ($CHR,$SNP,$ZERO,$POS) = split;
    $ZERO = "0";
    if ($SNP =~ "---") {
        $SNP = "$chr:$POS";
        if (exists $snp_hash{$SNP}) {
            $snp_hash{$SNP}++;
            $SNP = $SNP . '.' . $snp_hash{$SNP};
        }
    } else {
         if (exists $snp_hash{$SNP}) {
             $snp_hash{$SNP}++;
             $newSNP = $SNP . '.' . $snp_hash{$SNP};
             $SNP = $newSNP;
         }
    }
    $snp_hash{$SNP}++;
    print "$chr\t$SNP\t$ZERO\t$POS\n";
}

Do you try gtool? It can convert Impute2 output too PED/MAP format.

gtool -G --g file1 --s file2.sample --ped file3.ped --map file4.map --phenotype phenotype_1 --threshold 0.95 > output.gtool

---EDIT---

chr=1

awk -v var1=$chr '{
ORS = ""
print var1"\t"
if ($2 == "---") print "SNP."var1"."$4"\t"
else print $2"\t"
print $3"\t"
print $4"\n"
}' Chr${chr}.IMPUTE2.map > Chr${chr}.IMPUTE2.V2.map

I think this script is less complex. You can use it in a for loop with each chromosome in a separated file.

Thanks for the comment.

Also take a look at this python implementation:

http://www.pypedia.com/index.php/Convert_impute2_gprobs_to_PEDMAP_beagle_user_Kantale

From the long parameter list, you can only define the following parameters:

chromosome, input_impute2_gprobs_filename, input_impute2_info_filename, output_TPED_filename, output_TFAM_filename

Note: The output format is transposed PED/MAP files. You can use these files directly in plink with the --tfile parameter: http://pngu.mgh.harvard.edu/~purcell/plink/data.shtml#tr