Extract read names from SAM/BAM

get_names_sam.sh

#!/bin/bash
#
# Quickly extract unique read names from SAM/BAM file
# Source: https://www.biostars.org/p/371705/#371748 
#
# This will extract unique read names for all unampped reads (-f 4) from BAM file
# awk is much faster than `sort --parallel=$threads | uniq` because it doesn't have to do the sorting
# On BAM file with 149,909,118 input reads:
# sort | uniq takes (with 12 threads):
#real    1m8.240s
#user    2m57.097s
#sys     0m18.001s
# awk '!x[$0]++' takes (with 1 thread):
#real    0m42.857s
#user    2m30.028s
#sys     0m18.179s
#

threads=12
inbam="in.bam"

# sort | uniq version
#samtools view -@ $threads -f4 $inbam \
#    | cut -f1 | sort --parallel=$threads -T $(dirname $inbam) | uniq \
#    > out.unmapped-names.txt

# awk version
samtools view -@ $threads -f4 $inbam \
    | cut -f 1 | awk '!x[$0]++' \
    > out.unmapped-names.txt