Reputation: 13
I have a data set where ID's have 2 different occurences on the same day. There are about 10 different occurences. I want to cross tabulate the occurences using proc freq or proc tabulate & find how many times each instance occurs on the same day. I want my table to look something like this
Frequency occ1 occ2 occ3 occ4 occ5 occ6
occ1 2 0 0 1 4 0
occ2 1 0 0 0 0 0
occ3 3 0 0 0 0 0
occ4 0 5 3 0 3 0
occ5 0 2 4 0 5 0
occ6 1 5 4 2 1 2
My data looks something like this
data have;
input id occurrence ;
datalines;
id1 occ3
id1 occ2
id2 occ1
id2 occ6
id3 occ2
id3 occ4
etc...
i tried
proc freq data=have;
tables occurrence*occurence ;
run;
but not having any luck. I have tried other variations & using by ID but it gives every single ID individually & i have about 200 ID numbers.
Thanks!
Upvotes: 0
Views: 62
Reputation: 27498
Reform data so a tabulation of ordered pairs can be done.
data have;
call streaminit(2022);
do id = 1 to 20;
topic = rand('integer', 10); output;
topic = rand('integer', 10); output;
end;
run;
data stage;
do until (last.id);
set have;
by id;
row = col;
col = topic;
end;
run;
ods html file='pairfreq.html';
title "Ordered pair counts";
proc tabulate data=stage;
class row col;
table row='1st topic in id pair',col='2nd topic in id pair'*n='';
run;
ods html close;
Upvotes: 1