nPath Filters Example - Teradata Vantage

Machine Learning Engine Analytic Function Reference

Product
Teradata Vantage
Release Number
8.00
1.0
Published
May 2019
Language
English (United States)
Last Update
2019-11-22
dita:mapPath
blj1506016597986.ditamap
dita:ditavalPath
blj1506016597986.ditaval
dita:id
B700-4003
lifecycle
previous
Product Category
Teradata Vantageā„¢

Using clickstream data from an online store, this example finds the sessions where the user visited the checkout page within 10 minutes of visiting the home page. Because there is no way to know in advance how many rows might appear between the home page and the checkout page, the example cannot use a LAG or LEAD expression. Therefore, it uses the Filter argument.

Input

clickstream
userid sessionid clicktime pagetype
1 1 10-10-2012 10:15 home
1 1 10-10-2012 10:16 view
1 1 10-10-2012 10:17 view
1 1 10-10-2012 10:20 checkout
1 1 10-10-2012 10:30 checkout
1 1 10-10-2012 10:35 view
1 1 10-10-2012 10:45 view
2 2 10-10-2012 13:15 home
2 2 10-10-2012 13:16 view
2 2 10-10-2012 13:43 checkout
2 2 10-10-2012 13:35 view
2 2 10-10-2012 13:45 view

SQL Call

SELECT * FROM nPath@coprocessor (
  ON clickstream PARTITION BY userid ORDER BY clicktime
  USING
  Symbols (pagetype='home' AS home,
    pagetype!='home' AND pagetype!='checkout' AS view,
    pagetype='checkout' AS checkout)
  Pattern ('home.view*.checkout')
  Result (FIRST(userid of ANY(home, checkout, view)) AS userid,
    FIRST (sessionid of ANY(home, checkout, view)) AS sessionid,
    COUNT (* of any(home, checkout, view)) AS cnt,
    FIRST (clicktime of ANY(home)) AS firsthome,
    LAST (clicktime of ANY(checkout)) AS lastcheckout)
  Filter (FIRST (clicktime + '10 minutes' ::interval OF ANY (home)) > 
    FIRST (clicktime of any(checkout))) 
  Mode (NONOVERLAPPING)
) AS dt;

Output

userid sessionid cnt firsthome lastcheckout
1 1 4 2012-10-10 10:15:00 2012-10-10 10:20:00