Sorting structured file with bash commands

Question

I need to sort a LDIF file, where several lines belonging to a parent one.

Example

dn: 2

attr1: b

attr2: a

attr1: a

attr1: c



dn: 3

attr2: a

attr1: c

attr1: b

attr1: a



dn: 1

attr1: a

attr1: c

attr1: b

attr2: a

to this one

dn: 1

attr1: a

attr1: b

attr1: c

attr2: a


dn: 2

attr1: a

attr1: b

attr1: c

attr2: a


dn: 3

attr1: a

attr1: b

attr1: c

attr2: a

So all parent line starting with dn are sorted, below all attrx are sorted and if attrx has multi values there are also sorted. I have done this with read line, but this takes hours on huge files. Is there a faster way to do same with bash commands?

An attribute value is always taking only one line. If there are multiples values each take one line.No line is base64 encoded.

Answer 1

using your example file

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file |awk -F'*' '{for(i=1; i<=NF; i++) c[i]=$i; n=asort(c); for (i=1; i<=n; i++) printf "%s%s*", c[i], (i<n?OFS:RS); delete c}' |sed 's/^*//' |awk -F'*' '{print $5"*"$1"*"$2"*"$3"*"$4}' |sort |awk -F'*' 'BEGIN{OFS="\n\n";ORS="\n\n\n"} {print $1,$2,$3,$4,$5;}'

Convert every block of text to rows and using '*' to separate the fields

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file dn: 2*attr1: b*attr2: a*attr1: a*attr1: c dn: 3*attr2: a*attr1: c*attr1: b*attr1: a dn: 1*attr1: a*attr1: c*attr1: b*attr2: a

Sort fields within the lines and using '*' to separate the fields

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file |awk -F'*' '{for(i=1; i<=NF; i++) c[i]=$i; n=asort(c); for (i=1; i<=n; i++) printf "%s%s*", c[i], (i<n?OFS:RS); delete c}' |sed 's/^*//'

attr1: a *attr1: b *attr1: c *attr2: a *dn: 2 attr1: a *attr1: b *attr1: c *attr2: a *dn: 3 attr1: a *attr1: b *attr1: c *attr2: a *dn: 1

rearrange the fieslds in the lines to 'print dn: x' in first place

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file |awk -F'*' '{for(i=1; i<=NF; i++) c[i]=$i; n=asort(c); for (i=1; i<=n; i++) printf "%s%s*", c[i], (i<n?OFS:RS); delete c}' |sed 's/^*//' |awk -F'*' '{print $5"*"$1"*"$2"*"$3"*"$4}'

dn: 2*attr1: a *attr1: b *attr1: c *attr2: a dn: 3*attr1: a *attr1: b *attr1: c *attr2: a dn: 1*attr1: a *attr1: b *attr1: c *attr2: a

Sorting the lines by the first column or field

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file |awk -F'*' '{for(i=1; i<=NF; i++) c[i]=$i; n=asort(c); for (i=1; i<=n; i++) printf "%s%s*", c[i], (i<n?OFS:RS); delete c}' |sed 's/^*//' |awk -F'*' '{print $5"*"$1"*"$2"*"$3"*"$4}' |sort

dn: 1*attr1: a *attr1: b *attr1: c *attr2: a dn: 2*attr1: a *attr1: b *attr1: c *attr2: a dn: 3*attr1: a *attr1: b *attr1: c *attr2: a

Converting the rows to one column and inserting empty lines

awk 'BEGIN {RS="\n\n\n";FS="\n\n";OFS="*";ORS=""} {print $1,$2,$3,$4,$5}' file |awk -F'*' '{for(i=1; i<=NF; i++) c[i]=$i; n=asort(c); for (i=1; i<=n; i++) printf "%s%s*", c[i], (i<n?OFS:RS); delete c}' |sed 's/^*//' |awk -F'*' '{print $5"*"$1"*"$2"*"$3"*"$4}' |sort |awk -F'*' 'BEGIN{OFS="\n\n";ORS="\n\n\n"} {print $1,$2,$3,$4,$5;}'

dn: 1

attr1: a

attr1: b

attr1: c

attr2: a


dn: 2

attr1: a

attr1: b

attr1: c

attr2: a


dn: 3

attr1: a

attr1: b

attr1: c

attr2: a

I know I'm using too much steps.