No Scala book would be complete without mentioning the match/case statements. Scala has a very rich pattern-matching mechanism. For instance, let's say we want to find all instances of a sequence of page views that start with a homepage followed by a products page—we really want to filter out the determined buyers. This may be accomplished with a new function, as follows:
scala> def findAllMatchedSessions(h: Seq[Session[PageView]], s: Session[PageView]) : Seq[Session[PageView]] = { | def matchSessions(h: Seq[Session[PageView]], id: String, p: Seq[PageView]) : Seq[Session[PageView]] = { | p match { | case Nil => Nil | case PageView(ts1, "mycompanycom>homepage") :: PageView(ts2, "mycompanycom>plus>products landing") :: tail => | matchSessions(h, id, tail).+:(new Session(id, p)) | case _ => matchSessions(h, id, p.tail) | } | } | matchSessions(h, s.id, s.visits) | } findAllSessions: (h: Seq[Session[PageView]], s: Session[PageView])Seq[Session[PageView]]
Note that we explicitly put PageView constructors in the case statement! Scala will traverse the visits sequence and generate new sessions that match the specified two PageViews, as follows:
scala> rdd.flatMap(x => findAllMatchedSessions(Nil, x)).take(10).foreach(println) (88.234.248.111 -> [(2015-08-23 22:38:35 :mycompanycom>homepage),(2015-08-23 22:38:37 :mycompanycom>plus>products landing),(2015-08-23 22:39:01 :mycompanycom>homepage),(2015-08-23 22:39:24 :mycompanycom>homepage),(2015-08-23 22:39:26 :mycompanycom>plus>whatismycompanyfuel)]) (148.246.218.251 -> [(2015-08-23 22:52:09 :mycompanycom>homepage),(2015-08-23 22:52:16 :mycompanycom>plus>products landing),(2015-08-23 22:52:23 :mycompanycom>homepage),(2015-08-23 22:52:32 :mycompanycom>homepage),(2015-08-23 22:52:39 :mycompanycom>running:slp)]) (86.30.116.229 -> [(2015-08-23 23:15:00 :mycompanycom>homepage),(2015-08-23 23:15:02 :mycompanycom>plus>products landing),(2015-08-23 23:15:12 :mycompanycom>plus>products landing),(2015-08-23 23:15:18 :mycompanycom>language tunnel>load),(2015-08-23 23:15:23 :mycompanycom>language tunnel>geo selected),(2015-08-23 23:15:24 :mycompanycom>homepage),(2015-08-23 23:15:27 :mycompanycom>homepage),(2015-08-23 23:15:30 :mycompanycom>basketball:slp),(2015-08-23 23:15:38 :mycompanycom>basketball>lebron-10:cdp),(2015-08-23 23:15:50 :mycompanycom>basketball>lebron-10:cdp),(2015-08-23 23:16:05 :mycompanycom>homepage),(2015-08-23 23:16:09 :mycompanycom>homepage),(2015-08-23 23:16:11 :mycompanycom>basketball:slp),(2015-08-23 23:16:29 :mycompanycom>onsite search>results found),(2015-08-23 23:16:39 :mycompanycom>onsite search>no results)]) (204.237.0.130 -> [(2015-08-23 23:26:23 :mycompanycom>homepage),(2015-08-23 23:26:27 :mycompanycom>plus>products landing),(2015-08-23 23:26:35 :mycompanycom>plus>fuelband activity>summary>wk)]) (97.82.221.34 -> [(2015-08-23 22:36:24 :mycompanycom>homepage),(2015-08-23 22:36:32 :mycompanycom>plus>products landing),(2015-08-23 22:37:09 :mycompanycom>plus>plus activity>summary>wk),(2015-08-23 22:37:39 :mycompanycom>plus>products landing),(2015-08-23 22:44:17 :mycompanycom>plus>home),(2015-08-23 22:44:33 :mycompanycom>plus>home),(2015-08-23 22:44:34 :mycompanycom>plus>home),(2015-08-23 22:44:36 :mycompanycom>plus>home),(2015-08-23 22:44:43 :mycompanycom>plus>home)]) (24.230.204.72 -> [(2015-08-23 22:49:58 :mycompanycom>homepage),(2015-08-23 22:50:00 :mycompanycom>plus>products landing),(2015-08-23 22:50:30 :mycompanycom>homepage),(2015-08-23 22:50:38 :mycompanycom>homepage),(2015-08-23 22:50:41 :mycompanycom>training:cdp),(2015-08-23 22:51:56 :mycompanycom>training:cdp),(2015-08-23 22:51:59 :mycompanycom>store locator>start),(2015-08-23 22:52:28 :mycompanycom>store locator>landing)]) (62.248.72.18 -> [(2015-08-23 23:14:27 :mycompanycom>homepage),(2015-08-23 23:14:30 :mycompanycom>plus>products landing),(2015-08-23 23:14:33 :mycompanycom>plus>products landing),(2015-08-23 23:14:40 :mycompanycom>plus>products landing),(2015-08-23 23:14:47 :mycompanycom>store homepage),(2015-08-23 23:14:50 :mycompanycom>store homepage),(2015-08-23 23:14:55 :mycompanycom>men:clp),(2015-08-23 23:15:08 :mycompanycom>men:clp),(2015-08-23 23:15:15 :mycompanycom>men:clp),(2015-08-23 23:15:16 :mycompanycom>men:clp),(2015-08-23 23:15:24 :mycompanycom>men>sportswear:standardgrid),(2015-08-23 23:15:41 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:15:45 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:15:45 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:15:49 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:15:50 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:15:56 :mycompanycom>men>sportswear:standardgrid),(2015-08-23 23:18:41 :mycompanycom>pdp>mycompany bruin low men's shoe),(2015-08-23 23:18:42 :mycompanycom>pdp>mycompany bruin low men's shoe),(2015-08-23 23:18:53 :mycompanycom>pdp>mycompany bruin low men's shoe),(2015-08-23 23:18:55 :mycompanycom>pdp>mycompany bruin low men's shoe),(2015-08-23 23:18:57 :mycompanycom>pdp>mycompany bruin low men's shoe),(2015-08-23 23:19:04 :mycompanycom>men>sportswear:standardgrid),(2015-08-23 23:20:12 :mycompanycom>men>sportswear>silver:standardgrid),(2015-08-23 23:28:20 :mycompanycom>onsite search>no results),(2015-08-23 23:28:33 :mycompanycom>onsite search>no results),(2015-08-23 23:28:36 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:28:40 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:28:41 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:28:43 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:28:43 :mycompanycom>pdp>mycompany blazer low premium vintage suede men's shoe),(2015-08-23 23:29:00 :mycompanycom>pdp:mycompanyid>mycompany blazer low id shoe)]) (46.5.127.21 -> [(2015-08-23 22:58:00 :mycompanycom>homepage),(2015-08-23 22:58:01 :mycompanycom>plus>products landing)]) (200.45.228.1 -> [(2015-08-23 23:07:33 :mycompanycom>homepage),(2015-08-23 23:07:39 :mycompanycom>plus>products landing),(2015-08-23 23:07:42 :mycompanycom>plus>products landing),(2015-08-23 23:07:45 :mycompanycom>language tunnel>load),(2015-08-23 23:07:59 :mycompanycom>homepage),(2015-08-23 23:08:15 :mycompanycom>homepage),(2015-08-23 23:08:26 :mycompanycom>onsite search>results found),(2015-08-23 23:08:43 :mycompanycom>onsite search>no results),(2015-08-23 23:08:49 :mycompanycom>onsite search>results found),(2015-08-23 23:08:53 :mycompanycom>language tunnel>load),(2015-08-23 23:08:55 :mycompanycom>plus>products landing),(2015-08-23 23:09:04 :mycompanycom>homepage),(2015-08-23 23:11:34 :mycompanycom>running:slp)]) (37.78.203.213 -> [(2015-08-23 23:18:10 :mycompanycom>homepage),(2015-08-23 23:18:12 :mycompanycom>plus>products landing),(2015-08-23 23:18:14 :mycompanycom>plus>products landing),(2015-08-23 23:18:22 :mycompanycom>plus>products landing),(2015-08-23 23:18:25 :mycompanycom>store homepage),(2015-08-23 23:18:31 :mycompanycom>store homepage),(2015-08-23 23:18:34 :mycompanycom>men:clp),(2015-08-23 23:18:50 :mycompanycom>store homepage),(2015-08-23 23:18:51 :mycompanycom>footwear:segmentedgrid),(2015-08-23 23:19:12 :mycompanycom>men>footwear:segmentedgrid),(2015-08-23 23:19:12 :mycompanycom>men>footwear:segmentedgrid),(2015-08-23 23:19:26 :mycompanycom>men>footwear>new releases:standardgrid),(2015-08-23 23:19:26 :mycompanycom>men>footwear>new releases:standardgrid),(2015-08-23 23:19:35 :mycompanycom>pdp>mycompany cheyenne 2015 men's shoe),(2015-08-23 23:19:40 :mycompanycom>men>footwear>new releases:standardgrid)])
I leave it to the reader to write a function that also filters only those sessions where the user spent less than 10 seconds before going to the products page. The epoch trait or the previously defined to the EpochSeconds function may be useful.
The match/case function can be also used for feature generation and return a vector of features over a session.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.