Create Rule

This service of API is used for operation on creating CEP rules for a tenant. A user can create a wide variety of CEP rules with joining,filtering,aggregation etc functionalities and route the output to a topic.Input and output of CEP is topic.

Following is the URL for CEP create rule.

POST http://<domain name>/cep/v2.0/rules

CEP functionality can be classified into joining/filtering and pure aggregations.CEP rules can be wired and output of one CEP rule can be fed to another CEP rule. Classification of CEP rules is provided below :-

• MultipleStreamJoining/filtering/SQL on streaming data
• PureAggregation with Filtering

Following is sample json for creating rule :-

{
     "ruleName": "JoiningSample1",
     "selectClause": "s1.temp as currtemp,s2.humidity as currhumidity,s1.feature as feature,s1.starttime as s1time,s2.starttime as s2time",
     "joinClause": "(s1.feature = s2.feature) AND abs(s1.starttime - s2.starttime) <= 60",
     "whereClause": "currtemp > 40 AND currhumidity > 90",
     "groupByField": "feature",
     "triggerInterval": "30",
     "windowDuration": "",
     "input": [{
                     "alias": "s1",
                     "topic": "test1",
                     "properties": [
                             "temp"
                     ]
             },
             {
                     "alias": "s2",
                     "topic": "test2",
                     "properties": [
                             "humidity"
                     ]
             }
     ],
     "output": "outputjoin4"
}

Description of create rule json parameters is defined in below table

Parameter	Required	Description
ruleName	true	name of CEP rule,this has to be unique for the rules across tenants
selectClause	true	SQL interface for user, standard sql select query along with some supported spark sql specific select query with the help of keywords described below
whereClause	false	Condition clause with the alias or field already defined in “selectClause”, Example: tempdiff <= 0
joinClause	false	Join condition between two/many streams,based on common field,difference of event time in two streams along with optional filtering condition, Example: (s1.feature = s2.feature) AND abs(s1.starttime - s2.starttime) <= 60 AND (s1.temp > 40) AND (s2.humidity > 90)
groupByField	false	Similar to sql groupBy,output will be grouped based on the values in feature/sensor field in the input data
triggerInterval	false	User specified interval at which output will be triggered ,Example: 30(seconds)
windowDuration	false	Duration of window in seconds during joining or aggregation operations, Example: 30(seconds)
slidingDuration	false	Duration of slidingInterval in seconds(applicable only for purely aggregation functions), Example: 10(seconds)
input	true	alias - alias/streamname to uniquely identify topic topic - input topics from where cep rule will consume the data before processing properties - property/properties to be selected from the input topic
output	true	output topic where the output after processing will be written into

Here is the list of the TCUP CEP keyword, that can be used in the selectClause

Keyword	Description
rank	It sorts the fields by descending order of the values of the fields and assign rank to each field after sorting within the window.
lag(nth)	It returns the immediate previous nth value of the field within the window. lag(1) means immediate previous i.e n start with 1.
lead(nth)	It returns the immediate next nth value of the field within the window. lead(1) means immediate next i.e n start with 1.