summarizing a text file in awk - awk

Question

I have a sequence of characters in which I would like to split each sequence into 3-characters class from the beginning to the end. and the get the count of each class. here is a small example of sequences of characters for 2 IDs.
>ID1
ATGTCCAAGGGGATCCTGCAGGTGCATCCTCCGATCTGCGACTGCCCGGGCTGCCGAATA
TCCTCCCCGGTGAACCGGGGGCGGCTGGCAGACAAGAGGACAGTCGCCCTGCCTGCCGCC
>ID2
ATGAAACTTTCACCTGCGCTCCCGGGAACAGTTTCTGCTCGGACTCCTGATCGTTCACCT
CCCTGTTTTCCCGACAGCGAGGACTGTCTTTTCCAACCCGACATGGATGTGCTCCCAATG
ACCTGCCCGCCACCACCAGTTCCAAAGTTTGCACTCCTTAAGGATTATAGGCCTTCAGCT
and here is a small example of output for ID1. I want to get the same output for all IDs in the input file (the lines of characters belong each ID is in the next line). the counts for the next ID comes just after the first and so on.
ID1_3nt count
ATG 1
TCC 3
AAG 2
GGG 2
ATC 2
CTG 3
CAG 1
GTG 2
CAT 1
CCT 2
CCG 3
TGC 3
GAC 2
GGC 1
CGA 1
ATA 1
AAC 1
CGG 2
GCA 1
AGG 1
GCC 3
ACA 1
GTC 1
I tried this code:
awk '{i=0; printf ">%s\n",$2; while(i<=length($1)) {printf "%s\n", substr($1,i,3);i+=3}} /,substr,/ {count++}' | awk 'END { printf(" ID_3nt: %d",count)}
but did not return what I want. do you know how to improve it?

Answer 1

How about this patsplit()-based implementation?
#! /usr/bin/awk -f
# initialize publicly scoped vars...
function init() {
split("", idx) # index of our class (for ordering)
split("", cls) # our class name
split("", cnt) # num of classes we have seen
sz = 0 # number of classes for this ID
}
# process a class record
function proc( i, n, x) {
# split on each 3 characters
n = patsplit($0, a, /.../)
for (i=1; i<=n; ++i) {
x = a[i]
if (x in idx) {
# if this cls exists, just increment the count
++cnt[idx[x]]
} else {
# if this cls doesn't exist, index it in
cls[sz] = x
cnt[sz] = 1
idx[x] = sz++
}
}
}
# spit out class summary
function flush( i) {
if(!sz)
return
for(i=0; i<sz; ++i)
print cls[i], cnt[i]
init()
}
BEGIN {
init()
}
/^>ID/ {
flush()
sub(/^>/, "")
print $0 "_3nt count"
next
}
{
# we could have just inlined proc(), but using a function
# provides us with locally scoped variables
proc()
}
END {
flush()
}

Answer 2

$ cat tst.awk
sub(/^>/,"") { if (NR>1) prt(); name=$0; next }
{ rec = rec $0 }
END { prt() }
function prt( cnt, class) {
while ( rec != "" ) {
cnt[substr(rec,1,3)]++
rec = substr(rec,4)
}
print name "_3nt count"
for (class in cnt) {
print class, cnt[class]
}
}
.
$ awk -f tst.awk file
ID1_3nt count
ACA 1
AAC 1
CGA 1
CAT 1
GTG 2
CAG 1
GGG 2
CCG 3
CCT 2
GCA 1
ATA 1
GAC 2
AAG 2
GCC 3
ATC 2
TCC 3
CGG 2
CTG 3
GTC 1
AGG 1
GGC 1
TGC 3
ATG 1
ID2_3nt count
AAA 1
CCC 3
ACA 1
GTG 1
TTT 2
TGT 2
GTT 2
ACC 1
CCG 2
CTC 3
CCT 4
GCA 1
AAG 2
GAC 3
TCA 3
AGC 1
ACT 1
CGT 1
CGG 1
CTT 3
TAT 1
CAA 1
GAG 1
GAT 3
GGA 1
AGG 1
TGC 1
CCA 5
TTC 1
GCT 2
TCT 1
GCG 1
ATG 3