awk to extract value in each line and create new file - awk

Question

In the below awk I am trying to extract the value of a substring in each line, and the 2 attempts do not produce the desired results. The first awk executes and returns no data,
and the second only extracts the value. Thank you :).
file
#CHROM POS ID REF ALT QUAL FILTER INFO
1 930215 CM1613956 A G . . PHEN="Retinitis_pigmentosa";RANKSCORE=0.21
awk 1
awk '/^#/ {for (I=1;I<NF;I++) if ($I == "RANKSCORE=") print $(I+1)}' file
awk 2
awk 'BEGIN{FS=OFS="\t"}; /^#/ {print $1,$2,$3} {sub(/.*RANKSCORE=/, ""); print}' file
#CHROM POS ID
#CHROM POS ID REF ALT QUAL FILTER INFO
0.21
0.99
desired (tab-delimited)
1 930215 CM1613956 A G . . 0.21

Answer 1

You may use this awk:
awk 'BEGIN {FS=OFS="\t"}
/^#/ {next}
$NF ~ /;RANKSCORE=/ {
sub(/.+=/, "", $NF)
} 1' file
1 930215 CM1613956 A G . . 0.21

Answer 2

With your shown samples please try following awk code.
awk -F';RANKSCORE=' '
BEGIN{ OFS ="\t" }
/^#/ { next }
NF==2 && match($0,/.* /){
print substr($0,RSTART,RLENGTH-1),$2
}
' Input_file
Explanation: Adding detailed explanation for above code.
awk -F';RANKSCORE=' ' ##Starting awk program from here, settings field separator as ;RANKSCORE=
BEGIN{ OFS ="\t" } ##Setting OFS as tab in BEGIN section of this code.
/^#/ { next } ##If a line starts from # then simply skip that line.
NF==2 && match($0,/.* /){ ##Check if NF is 2 AND matching till last occurrence of single space.
print substr($0,RSTART,RLENGTH-1),$2 ##Printing sub string till matched regex along with 2md field.
}
' Input_file ##Mentioning Input_file name here.

Answer 3

Your RANKSCORE seems to appear in field 8.
match can locate it. substr can extract it.
$ awk -F'\t' -v OFS='\t' '
match($8,/RANKSCORE=[0-9.]+/){
$8 = substr($8, RSTART+10, RLENGTH-10)
print
}
' file
Or more safely, assuming semi-colon sub-delimiters, a couple of subs:
$ awk -F'\t' -v OFS='\t' '
sub(/^(.*;)?RANKSCORE=/,"",$8){
sub(/[^0-9.].*$/,"",$8)
print
}
' file

Answer 4

Assumptions:
we only want exact word matches on RANKSCORE (eg, do not match on old_RANKSCORE)
RANKSCORE=value could show up anywhere in a ;-delimited last field
Adding some lines with different locations of RANKSCORE:
#CHROM POS ID REF ALT QUAL FILTER INFO
1 930215 CM1613956 A G . . PHEN="Retinitis_pigmentosa";RANKSCORE=0.21
1 930215 CM1613956 A G . . RANKSCORE=3.235;PHEN="Retinitis_pigmentosa"
1 930215 CM1613956 A G . . stuff=123;old_RANKSCORE=7.7234;PHEN="Retinitis_pigmentosa"
1 930215 CM1613956 A G . . stuff=123;RANKSCORE=9.3325;PHEN="Retinitis_pigmentosa"
One awk idea:
awk '
BEGIN { FS=OFS="\t" }
/RANKSCORE/ { n=split($NF,a,"[;=]") # if line contains "RANKSCORE" then split last field on dual delimiters ";" and "="
for (i=1;i<=n;i=i+2) # loop through attribute names (odd-numbered indices) and ...
if (a[i] == "RANKSCORE") { # if attribute == "RANKSCORE" then ...
$NF=a[i+1] # use associated value (even-numbered index) as new value for last field
print # print new line
next # go to next input line
}
}
' file
This generates:
1 930215 CM1613956 A G . . 0.21
1 930215 CM1613956 A G . . 3.235
1 930215 CM1613956 A G . . 9.3325

Answer 5

no arrays needed :
{m,n,g}awk '!+_<+NF && sub(";.*$", _, $(NF=NF))^_'\
FS='[ \t]+([^ \t]*;)?RANKSCORE=' OFS='\t'
1 930215 CM1613956 A G . . 0.21
1 930215 CM1613956 A G . . 3.235
1 930215 CM1613956 A G . . 9.3325