sort with header maintained

We usually want to maintain header when we do something for the body of the text.
You can find a body function here which can only accept streams and assumes only one line header.

Use subshells and awk can do the same job and potentially more flexible.

	## imagine we have a file with one line header, and we want to keep the header after sorting
	## use subshells http://bash.cyberciti.biz/guide/What_is_a_Subshell%3F
	(sed -n '1p' your_file; cat your_file | sed '1d' | sort) > sort_header.txt
	## if you have two header lines and want to keep both of them:
	(sed -n '1,2p' your_file; cat your_file | sed '1,2d' | sort) > sort_header.txt
	## if you have many lines starting with "#" as header, like vcf files
	(grep "^#" my_vcf; grep -v "^#" my_vcf | sort -k1,1V -k2,2n) > sorted.vcf
	## one can also use awk
	cat my_vcf | awk '$0~"^#" { print $0; next } { print $0 | "LC_ALL=C sort -k1,1V -k2,2n" }'
	## I am a useless cat user :) http://stackoverflow.com/questions/11710552/useless-use-of-cat

view raw maintain_header_sort.sh hosted with ❤ by GitHub

The original credits go to Aaron Quinlan. see a gist below to sort vcf files in natural chromosome order :chr1 chr2 chr3.... rather than chr1 chr10 chr11....

	chmod a+x vcfsort.sh
	vcfsort.sh trio.trim.vep.vcf.gz

view raw natural_sort_vcf.sh hosted with ❤ by GitHub

	sort VCF and keep (only the first) header as-is:
	awk 'BEGIN{x=0;} $0 ~/^#/{ if(x==0) {print;} next}{x=1; print $0 | "sort -k1,1 -k2,2n"}'

view raw sort_vcf_keep_header.sh hosted with ❤ by GitHub

	#!/bin/bash
	# Faster, but can't handle streams
	[ $# -eq 0 ] && { echo "Sorts a VCF file in natural chromosome order";\
	echo "Usage: $0 [my.vcf | my.vcf.gz]"; exit 1;
	}
	# cheers, @michaelhoffman
	if (zless $1 | grep ^#; zless $1 | grep -v ^# | LC_ALL=C sort -k1,1V -k2,2n);
	then
	exit 0
	else
	printf 'sort failed. Does your version of sort support the -V option?\n'
	printf 'If not, you should update sort with the latest from GNU coreutils:\n'
	printf 'git clone git://git.sv.gnu.org/coreutils'
	fi

view raw vcfsort.sh hosted with ❤ by GitHub

	#!/bin/bash
	# Slower, but handles streams.
	[ $# -eq 0 ] && { echo "Sorts a VCF file in natural chromosome order";\
	echo "Usage: $0 [my.vcf | my.vcf.gz]"; \
	echo "Usage: cat [my.vcf | my.vcf.gz] | $0"; \
	exit 1;
	}
	# cheers, @colbychiang
	if zless $1 | awk '$0~"^#" { print $0; next } { print $0 | "LC_ALL=C sort -k1,1V -k2,2n" }';
	then
	exit 0
	else
	printf 'sort failed. Does your version of sort support the -V option?\n'
	printf 'If not, you should update sort with the latest from GNU coreutils:\n'
	printf 'git clone git://git.sv.gnu.org/coreutils'
	fi

view raw vcfsort2.sh hosted with ❤ by GitHub