Incorporate variables into bash code line

Question

Well maybe this is not the best title; but it's hard to convey my intention only in short title.

I've a line here:

2   118610455   P2_PM_2_5034    T   <DUP:TANDEM>    40  .   END=118610566;SVLEN=110;SVTYPE=TDUP;CIPOS=-100,55;CIEND=-56,100;IMPRECISE;DBVARID=esv7540;VALIDATED;VALMETHOD=CGH;SVMETHOD=RP

Basically I would like to convert it into:

2 118610455 118610566

So major problem is to grep this 118610566 from the 8th column.

I know how to grep this number:

$c=`cat line|awk '{print $8}'|sed 's/;/\t/g'|awk '{print $1}'|sed 's/\END=//g'`

but my question is then how I can incorporate this variable into another bash line:

what_i_want=`cat line|awk '{print $1"\t"$2"\t"$c}'`

thx

jaypal singh · Accepted Answer · 2011-12-19 22:51:46Z

3

May be this can help -

[jaypal:~/Temp] cat tmp
2   118610455   P2_PM_2_5034    T   <DUP:TANDEM>    40  .   END=118610566;SVLEN=110;SVTYPE=TDUP;CIPOS=-100,55;CIEND=-56,100;IMPRECISE;DBVARID=esv7540;VALIDATED;VALMETHOD=CGH;SVMETHOD=RP

[jaypal:~/Temp] var=$(awk -v FS="[ ;=]" '{print $1,$4,$24}' tmp)

[jaypal:~/Temp] echo $var
2 118610455 118610566

FS is awk's built-in variable. It is defaulted to a space or a tab. Since your line as more than one delimiter setting the FS to a character class helps in splitting the line for each de-limiter. The character class we have defined here is either a space, semi-colon or equal.

Might feel a little odd but I use this as a my debugging tool for identifying columns when I happen to parse a line with more than 1 delimiters. This is what I had got from your line -

[jaypal:~/Temp] awk -v FS="[ ;=]" '{for(i=1;i<=NF;i++) print "$"i" is "$i}' tmp
$1 is 2
$2 is 
$3 is 
$4 is 118610455
$5 is 
$6 is 
$7 is P2_PM_2_5034
$8 is 
$9 is 
$10 is 
$11 is T
$12 is 
$13 is 
$14 is <DUP:TANDEM>
$15 is 
$16 is 
$17 is 
$18 is 40
$19 is 
$20 is .
$21 is 
$22 is 
$23 is END
$24 is 118610566
$25 is SVLEN
$26 is 110
$27 is SVTYPE
$28 is TDUP
$29 is CIPOS
$30 is -100,55
$31 is CIEND
$32 is -56,100
$33 is IMPRECISE
$34 is DBVARID
$35 is esv7540
$36 is VALIDATED
$37 is VALMETHOD
$38 is CGH
$39 is SVMETHOD
$40 is RP

You can also use a simple substr built-in function of awk in the following manner -

[jaypal:~/Temp] awk '{print $1,$2,$8=substr($8,5,9)}' tmp
2 118610455 118610566

edited Dec 19, 2011 at 22:51

answered Dec 19, 2011 at 21:51

jaypal singh

77.6k24 gold badges108 silver badges147 bronze badges

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

1 Comment

LookIntoEast Over a year ago

thx but can you explain about FS="[;=]" a bit? I don't know why 11861045 becomes the 4th column.

Kevin · Accepted Answer · 2011-12-19 22:17:18Z

1

With a little string manipulation you can get it in one go.

what_i_want=$(awk '{sub(/^END=/,"",$8); sub(/;.*$/,"",$8); print $1,$2,$8}' line)

Some explanation:

sub(a,b,c) searches for pattern a in variable c and replaces it with b, storing the modified string back into c. Patterns are written within //.

^ is the beginning of the string, $ is the end, . is anything, and * means zero or more of the preceding pattern. So in our case:

sub(/^END=/,"",$8); matches END= at the beginning (^) of the string and replaces it with "", nothing, essentially deleting it.

sub(/;.*$/,"",$8); takes everything (.*) from ; to the end ($) and deletes it. Note that in awk, as with most regex engines, * is greedy, which means it takes the longest match it can get, so we know this will get the first ;.

And all we are left with is the number you want.

edited Dec 19, 2011 at 22:17

answered Dec 19, 2011 at 21:42

Kevin

56.6k15 gold badges107 silver badges139 bronze badges

1 Comment

LookIntoEast Over a year ago

thx works well...but can you explain a bit about sub(/;.*/,"",$8) ? I know here to truncate the part after ; right? but I don't understand what . and * means here.

ghoti · Accepted Answer · 2011-12-19 21:52:26Z

If your "columns" are always separated by whitespace, then you don't need to use subshells and awk, you can do this directly in shell:

[ghoti@pc ~]$ read one two three four five junk <<< "2   118610455   P2_PM_2_5034    T   <DUP:TANDEM>    40  .   END=118610566;SVLEN=110;SVTYPE=TDUP;CIPOS=-100,55;CIEND=-56,100;IMPRECISE;DBVARID=esv7540;VALIDATED;VALMETHOD=CGH;SVMETHOD=RP"
[ghoti@pc ~]$ echo "$five"
<DUP:TANDEM>
[ghoti@pc ~]$ echo "$junk"
40 . END=118610566;SVLEN=110;SVTYPE=TDUP;CIPOS=-100,55;CIEND=-56,100;IMPRECISE;DBVARID=esv7540;VALIDATED;VALMETHOD=CGH;SVMETHOD=RP

The last variable specified on your read line gets "everything else".

Also. if you're handling multiple lines like this, you can run it in a loop:

cat /path/to/inputfile | while read one two three four five junk; do
  echo "$one - $two - $five"
done

Salt to taste.

Collectives™ on Stack Overflow

Incorporate variables into bash code line

3 Answers 3

1 Comment

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related