==> user.application.2020-01-16-00-00.csv

user1,app1
user1,app2
user2,app1
user3,app1

==> user.application.2020-01-16-00-30.csv

user1,app1
user2,app1
user2,app4
user10,app2
user10,app1
user4,app5

I want output like as follows, app followed by distinct number of users

app1,4
app2,2
app4,1
app5,1

CodePudding user response：

Store all combinations in an array and print the length of the array.

awk -F, '{a[$2][$1]} END { for (i in a) { print i "," length(a[i]) } }' *.csv