script to associate columns - arrays

Question

I have a file consisting of definitions where a variable name points to an ip address. I want a script (bash/python/similar) to output each ip address followed by a list of each variable where it is defined as well as how many variables were in the list.
Input:
define alpha 192.168.1.1
define beta 192.168.1.3
define gamma 192.168.1.2
define delta 192.168.1.1
define epsilon 192.168.1.3
define zeta 192.168.1.1
define eta 192.168.1.3
define theta 192.168.1.1
Output
192.168.1.1:alpha,delta,zeta,theta:4
192.168.1.3:beta,epsilon,eta:3
192.168.1.2:gamma:1
Do I use associative arrays in bash to do this or is there a better way? I tried to do it but only ended up with a bash script which I had to combine with linux sort and uniq commands but still couldn't get it quite right.
Apologies for crappy title but I couldn't formulate this in a better way so feel free to edit.

Answer 1

1st solution: Could you please try following. This should be able to have those values even you have more than 1 values after keyword define(1st field) and between last field(IP address), though your example do not have more than 3 fields but I have taken care of that(in case your Input_file is having more than 3 fields in it too).
awk '{$1="";val=$NF;$NF="";gsub(/^ +| +$/,"");a[val]=a[val]?a[val]","$0:$0;b[val]++} END{for(i in a){print i":"a[i]":"b[i]}}' Input_file
OR adding a non-one liner form of solution here.
awk '
{
$1=""
val=$NF
$NF=""
gsub(/^ +| +$/,"")
a[val]=a[val]?a[val]","$0:$0
b[val]++
}
END{
for(i in a){
print i":"a[i]":"b[i]
}
}' Input_file
2nd solution: Above 1st solution will NOT give same order of IPs(which they are present in Input_file), this solution will take care of sequence of IPs in output should be same as Input_file's presence.
awk '{$1="";val=$NF;$NF="";gsub(/^ +| +$/,"")} !c[val]++{d[++count]=val} {a[val]=a[val]?a[val]","$0:$0;b[val]++} END{for(i=1;i<=count;i++){print d[i]":"a[d[i]]":"b[d[i]]}}' Input_file
OR(adding a non-one liner form of above solution).
awk '
{
$1=""
val=$NF
$NF=""
gsub(/^ +| +$/,"")
}
!c[val]++{
d[++count]=val
}
{
a[val]=a[val]?a[val]","$0:$0
b[val]++
}
END{
for(i=1;i<=count;i++){
print d[i]":"a[d[i]]":"b[d[i]]
}
}' Input_file

Answer 2

Using awk:
awk '
/^define/{
a[$3]++
b[$3]=(b[$3]?b[$3]",":"")$2
}
END {
for(i in a)
print i,b[i],a[i]
}' OFS=: file
a is an array that holds the count of each different IP.
b is an array that holds a string containing all keywords for each IP. A comma is inserted in between every keyword.
At the end of the parsing the index containing the IP and both arrays content are printed.
The OFS=: sets the output field separator to :.