Finding Pattern

Vrushank — Mon, 02 Jul 2018 21:39:01 GMT

I have a dataset which has all the clicks done on the website in 1 column. I want to find the pattern which gets repeated in the whole data and the data contains more than 1 Million rows and has 17000 different pattern. I also want to know the average time spend on each click for each pattern. I have written a code in SAS which groups each pattern and also finds the time difference between each click but I am not getting the output how I want. For example, according to my code I am getting this output:

Clicks Group Time(Seconds)

A 1 6
B 1 2
C 1 0
D 2 12
E 2 5
F 2 0
A 3 9
B 3 6
C 3 0
H 4 8
I 4 9
J 4 0

Output expected:

Clicks AverageTime Count

ABC A-7.5,B-4,C-0 2
DEF D-12,E-5,F-0 1
HIJ H-8,I-9,J-0 1

Re: Finding Pattern

Astounding — Mon, 02 Jul 2018 23:17:16 GMT

Are we guaranteed that every group contains exactly 3 clicks?

Here's an approach, that does the heavy lifting. You can easily expand upon it if more than 3 clicks per pattern are allowed:

data want;

length pattern $ 100;

array duration {3};

recnum = 0;

do until last.group;

set have;

by group;

if first.group then pattern = clicks;

else pattern = catx('|', pattern, clicks);

recnum + 1;

duration{recnum} = time;

end;

keep group pattern duration1-duration3;

run;

This will at least give you the pieces to work with. The rest of the programming would use simple steps like PROC FREQ and PROC MEANS.

topic Finding Pattern in SAS Programming

Finding Pattern

Re: Finding Pattern