Bash array grouping elements

Question

I'm trying to create an array that looks like this:

0       kernel
1       /boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64 
2       kernel
3       /boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64

Using the grubby command of which the ouput for the entires I'm looking for are:

index=2
kernel=/boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64
root=/dev/mapper/<my_vol>-root
initrd=/boot/initramfs-4.8.0-1.el7.elrepo.x86_64.img
title=CentOS Linux (4.8.0-1.el7.elrepo.x86_64) 7 (Core)
index=3
kernel=/boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64
root=/dev/mapper/<my_vol>-root
initrd=/boot/initramfs-4.7.3-1.el7.elrepo.x86_64.img
title=CentOS Linux (4.7.3-1.el7.elrepo.x86_64) 7 (Core)

(I've removed the "args=" line for sanitisation purposes)

When I try to use the following command which works... (err... for the first 2 elements of the array). It decided that 2 indicies should be squashed into one.

IFS='=' read -a KERNELS <<< $(sudo grubby --info=ALL |grep -v rescue |grep -E 4.[78] |grep kernel)

0       kernel
1       /boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64 kernel
2       /boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64

First of why is it doing this? And how can I fix it?

I understand that there are other ways I could solve this problem but I want to understand what's going on here:

for i in ${KERNELS[@]}; do echo $i; done
kernel
/boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64
kernel
/boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64

for i in ${!KERNELS[@]}; do echo $i ${KERNELS[$i]}; done
0 kernel
1 /boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64 kernel
2 /boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64

I suspect it has something to do with the fact that when you run

sudo grubby --info=ALL |grep -v rescue |grep -E 4.[78] |grep kernel
kernel=/boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64
kernel=/boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64

You get two lines of output and the IFS= is separating on '=' and ignoring the '\n'

I've tried using awk '{print}' ORS=' ' ideas to output everything on one line but it gets messy.

There has to be a more elegant way to use the hereto string with multiple lines of output. Surely...

As an aside, the long grep pipeline seems to beg for an Awk solution. grubby --info=ALL | awk 'BEGIN { ORS="=" } !/rescue/ && /4\.[78]/ && /kernel/' -- notice also the requirement to backslash a literal period, and see also useless use of grep — tripleee
– tripleee, Commented Mar 7, 2017 at 12:37
Quote your parameter expansions: for i in ${!KERNELS[@]}; do echo "$i" "${KERNELS[i]}"; done. You'll see that ${KERNELS[1]} contains the newline that separates the two lines in the output of the pipeline. A here-string isn't appropriate here. (Also, what version of bash are you using? A bug dealing with pre-command environment settings and here-strings that wasn't completely fixed until bash 4.4 may be affecting your results.) — chepner
– chepner, Commented Mar 7, 2017 at 12:57

anubhava · Accepted Answer · 2017-03-07 11:32:39Z

1

You can use process substitution combined with read directive:

arr=()

while IFS='=' read -r k v; do
   [[ $k = "kernel" ]] && { arr+=("$k"); arr+=("$v"); }
done < <(sudo grubby --info=ALL)

Check array content:

printf '[%s]\n' "${arr[@]}"

[kernel]
[/boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64]
[kernel]
[/boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64]

answered Mar 7, 2017 at 11:32

anubhava

790k67 gold badges603 silver badges671 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

b0bu · Accepted Answer · 2017-03-07 11:53:43Z

0

Ok so I accidently just has a moment of clarity using the exta info posted above

IFS='=' read -a KERNELS <<< $(sudo grubby --info=ALL |grep -v rescue |grep -E 4.[78] |grep kernel | awk '{print}' ORS='=')

for i in ${!KERNELS[@]}; do echo ${KERNELS[$i]} $i; done
kernel 0
/boot/vmlinuz-4.8.0-1.el7.elrepo.x86_64 1
kernel 2
/boot/vmlinuz-4.7.3-1.el7.elrepo.x86_64 3

I've used awk to get around the fact that there are multiple lines with terminating '\n' and instead inserted a '=' which will be the delimiting factor in building the array. =]

answered Mar 7, 2017 at 11:53

b0bu

1,2781 gold badge11 silver badges31 bronze badges

3 Comments

anubhava Over a year ago

You don't really need to use 2 grep commands and one awk after grubby command to parse it out. Just a single awk can handle it all.

b0bu Over a year ago

I know this. But since unix pipes, unix programs, and the unix paradigm in general is based on small, simple pieces chained together, why would I replace 3 simple easily readable/editable stages for ascii vomit I'm going to have to groan at in 6 months. Also I have other problems to solve ^_^ but thanks for your input!

anubhava Over a year ago

It's your choice but in case your input is big than invoking 3 external calls will hurt you big time while processing.

Collectives™ on Stack Overflow

Bash array grouping elements

2 Answers 2

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related