[SOLVED] How to get search results to array by awk

webhope · 05-05-2010, 06:45 AM

I have this multiline variable:

Code:

block_new="title Sata Mandriva
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=eab515e9-bc3e-4024-9f01-55fddaa0fb1c  resume=UUID=e12487ff-6d6f-44c4-9e03-33f545b3b798 splash=silent vga=788
initrd (hd0,2)/boot/initrd.img"

In one thread here grail written this code to get array with results of search of uuids:

Code:

uuid=($(echo "$block_new" | awk 'BEGIN{RS="UUID="}/-/{gsub(/ .*/,"");print}'))

This returns array with uuids.

Now I tried several ways how to get similar array but with values like hd(0,2) and similar.

These are the tested codes

Code:

#1
hd=($(echo "$block_new" | awk '/-/{ x=gensub(/.*(\(hd[[:digit:]],[[:digit:]]+\)).*/,"\\1","g");  print x }'))
#2
hd=($(echo "$block_new" | awk 'BEGIN{RS="hd"}/-/{ x=gensub(/([[:digit:]],[[:digit:]]+)).*/,"\\1)","");print $0="(hd"x}'))
#3
hd=($(echo "$block_new" | awk 'BEGIN{RS="hd"}/-/{ x=gensub(/([[:digit:]],[[:digit:]]+)).*/,"(hd\\1)","");print x}')) 
#5
hd=($(echo "$block_new" | awk 'BEGIN{RS="hd"}/-/{ gsub(/([[:digit:]],[[:digit:]]+\)).*/,"(hd\1)"); print}'))
echo ${hd[@]};

The codes #1 to #4 should return result (hd0,2). Problem is that it is not a array.

Last try #5 more similar to grail's code. #5 returns "(hd)" without substitution. I run out of ideas. Some ideas?

David the H. · 05-05-2010, 08:27 AM

Using the sample you gave, all of your awk statements only produce a single text string.

The pattern "array=( )" will set one array element for each separate* string inside the parentheses. For example, "array=( abc def ghi )" will set three elements. Since awk is only outputting one string, there's only one entry.

I duplicated your text string 3 times, with slight modifications. When I did, your expressions 2-5 work (#1 doesn't match properly).

Code:

$ block_new='title Sata Mandriva
kernel (hd0,1)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=eab515e9-bc3e-4024-9f01-55fddaa0fb1c
resume=UUID=e12487ff-6d6f-44c4-9e03-33f545b3b798 splash=silent vga=788
initrd (hd0,1)/boot/initrd.img
title Sata Mandriva
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=eab515e9-bc3e-4024-9f01-55fddaa0fb1c
resume=UUID=e12487ff-6d6f-44c4-9e03-33f545b3b798 splash=silent vga=788
initrd (hd0,2)/boot/initrd.img
title Sata Mandriva
kernel (hd0,3)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=eab515e9-bc3e-4024-9f01-55fddaa0fb1c
resume=UUID=e12487ff-6d6f-44c4-9e03-33f545b3b798 splash=silent vga=788
initrd (hd0,3)/boot/initrd.img'


$ hd=( $(echo "$block_new" | awk 'BEGIN{RS="hd"}/-/{ x=gensub(/([[:digit:]],[[:digit:]]+)).*/,"\\1)","");print $0="(hd"x}') )

$ echo "${hd[0]}"
(hd0,1)
$ echo "${hd[1]}"
(hd0,2)
$ echo "${hd[2]}"
(hd0,3)

*What's considered a separate string depends on the shell' IFS (internal field seperator) setting. This is space/tab/newline by default.

grail · 05-05-2010, 08:41 AM

Code:

block_new="title Sata Mandriva
kernel (hd0,2)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=eab515e9-bc3e-4024-9f01-55fddaa0fb1c  resume=UUID=e12487ff-6d6f-44c4-9e03-33f545b3b798 splash=silent vga=788
initrd (hd0,2)/boot/initrd.img"

hd=$(echo "$block_new" | awk 'BEGIN{FS="\n";RS=""}{print gensub(/.*(\(.*\)).*/,"\\1","g")}')

echo $hd

Alternate "hd=" line:

Code:

hd=$(echo "$block_new" | awk 'gsub(/^k[^(]+|\/.*vm.*/,"")')

webhope · 05-05-2010, 11:38 AM

Quote:

Originally Posted by David the H.

I duplicated your text string 3 times, with slight modifications. When I did, your expressions 2-5 work (#1 doesn't match properly).

Oops! I thought that the result of my code should be:
$ echo "${hd[@]}"
(hd0,2) (hd0,2)

But you surprised me. I hadn't idea to change the number. That is exactly what I need. Because I don't want duplicate numbers/item/records in the array.

webhope · 05-05-2010, 11:54 AM

Quote:

Originally Posted by grail

Alternate "hd=" line:

Code:

hd=$(echo "$block_new" | awk 'gsub(/^k[^(]+|\/.*vm.*/,"")')

Also interesting. It looks that this is coming out from presumption that the source string is not dynamical. The substring to remove begins at "k" character (kernel). The second part to remove which is after pipeline specifies the word "vmlinuz". Right?

grail · 05-05-2010, 11:59 AM

Quote:

Also interesting. It looks that this is coming out from presumption that the source string is not dynamical. The substring to remove begins at "k" character (kernel). The second part to remove which is after pipeline specifies the word "vmlinuz". Right?

Correct