Data Flow rule form - Completing Data Flows

Use the source pattern to define data that you read from the data flow. A source is the standard entry point of a data flow.

• A data flow contains one primary source. This primary source can originate from:

  • Abstract – The connection point for other data flows (nested data flows). It can represent the input page to be used when invoking either Save or Process operations.
  • Data flow – The entry point is based on another data flow. This data flow can be under the same data flow class or another class.

    When you select a data flow as the source or destination of other data flows, you create data flow chains.

    Only a data flow with destination of type Abstract can be used as the primary input in another data flow.

  • Data set – A collection of data that is arranged in rows and columns for further processing, for example, data that contains customer interaction information.

    You can control whether the data set that is the source of the data flow behaves like a stream of data or not by checking or clearing the Use this data set as stream check box. However, depending on the type, data sets can be used only as streams (for example, Kafka or Stream), only as non-streams (for example, Database Table or File) or they can be used either as streams or non-streams (Monte Carlo only).

  • Report definition – The entry point is based on a report definition. The scope of this report definition is the Applies to class of the data flow. To distribute a data flow run across multiple nodes, select the Enable distributed run check box and specify the Partition Key.

    To use an aggregate report definition that generates a database group by statement, create a property in the class on which the report is defined. The property must have the same name as the aggregate in the report definition.

• Secondary sources are added as you start combining data. Secondary data sources can originate from a data set, data flow, or report definition.

Use the compose pattern to combine data from two sources into a page or page list property. This pattern requires a property to match data between the two sources. The starting data point is the shape from which you start combining data. After the compose path is available, you can define the secondary source by providing an existing data set or another data flow.

You create the convert pattern to change the class of the incoming data pages to another class. The following conversion modes are available:

• Top-level – Converts the top-level page.
• Embedded – Extracts and converts a property that is embedded in the top-level page list property. The type of the property can be Page or Page List. You can use this mode to propagate only part of the incoming data to a branched destination in a data flow. For example, you can propagate strategy results (without customer data) to the Interaction History data set. The page that is the source for the unpacked property can be preserved and propagated to another destination in the data flow through a different branch.

Class conversion involves overwriting identical properties by name, or manually mapping properties that do not have the same name between the source and target.

The conversion pattern requires a property to match data between the two sources. The data that you pass is determined by the starting shape for the convert path. After the convert path is available, you define its properties by selecting the target class and specifying how to handle property mapping between the source and target.

The mapping of properties between the source and target can be handled automatically, or you can manually define how to map source and target properties.

Configuring the convert pattern

1. In the Convert page on <input class name> section, select the conversion mode:
   • To convert the top-level page and all its embedded pages, perform the following actions:
     1. Select the Top-level check box.
     2. In the Into page(s) of class field, press the Down Arrow key and select the target class for the converted data.
   • To convert a page that is embedded in the top-level page, perform the following actions:
     1. Select the Embedded check box.
     2. In the Select page(list) field, press the Down Arrow key and select a page or page list property that is embedded in the top-level page.
2. Optional: Map the source class properties to target class properties:
   • To map properties automatically, in the Field mapping section, select the Auto-copy properties with identical names check box. Expand the drop-down list to view all properties whose names match in the source and target class. If you select this option, the properties in the source class overwrite properties with the same name in the target class.
   • To map properties manually, perform the following actions:
     1. Click Add mapping.
     2. In the Target column, press the Down Arrow key and select a property from the target class.
     3. In the Source column, press the Down Arrow key and select a property from the source class.

     This way, you create expressions, for example Set .pyName equal to .pySubjectName, and so on. Through manual mapping, you control which properties are overwritten because they do not have the same name but represent the same data.

3. Click Submit.

By adding this shape in a data flow, you can apply complex data transformations on the top-level clipboard page through data transform rules. In this shape, you can reference data transform rules that belong to the following classes:

• The Applies to class of the input data pages.
• A parent class of the Applies to class of the input data pages.

When a data transformation is complete, the Data Transform shape propagates the modified pages to the next shape in the flow.

Configuring the data transform pattern

1. In the Data transform field, press the Down Arrow key and select the data transform rule to apply on the incoming data.
2. Click Submit.

Use the event strategy pattern to reference and run an event strategy in a data flow.

Configuring the event strategy pattern

1. In the Results in field, enter the name of the class that contains your data flow.
2. In the Event strategy field, enter the name of the event strategy that you want to use.
3. In the Properties output mapping section, map the properties from your event strategy to the properties that exist in the class containing your data flow by performing the following steps:
1. Click Add Item.
2. In the Set field, enter a target property that is in the same class as your data flow.
3. In the Equal to field, enter a source property.

Use the filter pattern to define and apply a filter to the incoming data. The output consists of those elements that satisfy the filter conditions.

Configuring the filter pattern

1. In the Name field, enter a name for your filter.
2. In the Filter conditions section, click Add Item.

3. In the left field, enter the name of a property that is evaluated by the filter.

4. From the dropdown list, select a comparison operator.
5. In the right field, enter a value for your filter condition.

Use the Text Analyzer pattern to reference an instance of the Text Analyzer rule to apply text analysis on the incoming records that contain text.

Use the merge pattern to combine data in the primary and secondary data paths of a data flow into a single track. The starting data point is the shape from which you start merging data. When the merge path is available, you define the secondary source by providing an existing data set or another data flow.

In cases of no data match, you can exclude the source component results that do not match the merge condition. If one of the specified properties does not exist, the value of the other property is not included in the class that stores the merge results.

In cases of data mismatch, you can select which source takes precedence:

• Primary path – if there are properties with the same name but different values, merge takes the value in the primary source.
• Secondary path – if there are properties with the same name but different values, merge takes the value in the secondary source.

Use the strategy pattern to run a strategy based on combined or not combined data. The data that you pass to run the strategy is determined by the starting shape for the strategy path. When the strategy path is available, you select the strategy to run and the mode in which to run it. In a data flow, each strategy can run in one mode only.

You can use the following modes for running a strategy:

• Make decision - Use this mode when the strategy executed by the data flow is designed only to issue a decision.
• Make decision and store data for later response capture - Use this mode when the strategy executed by the data flow is designed to issue a decision and you want to store the decision results for a specified period of time. You can use this data for delayed adaptive model learning and issuing a response capture at a later time.
  • Store data for - Specify how long you want to store inputs passed to adaptive models and strategy results.
• Capture response for previous decision - Use this mode when the strategy executed by the data flow is designed to retrieve the adaptive inputs and strategy results. To retrieve decision results, you select one of the following options:
  • Identified by interaction ID - Enter the interaction ID for which you want to retrieve the adaptive inputs and strategy results.
  • In the past - Specify the time period from which you want to retrieve the adaptive inputs and strategy results.

Output strategy results

Unfold this section and select a class where you want to store strategy results.

• Individually in <class_name> - Use the strategy result class (default option). Each result is emitted to the destination individually.
• Updated in <class_name> - Use the input class of the strategy pattern as the output class. You can embed the strategy results in the top-level page.
• Embedded in - Enter any other class to store your strategy results. You can embed the strategy results in a different class.

When you change the default output class, map the properties from the strategy result class to the properties of the class that you select.

Use the destination pattern to define the data point you write to and it is the standard output point of a data flow. Every data flow defines one or more destinations that output all results, or results based on a given condition.

You can use the following destination types:

• Abstract
• Activity
• Case
• Data flow
• Data set

• When you select a data flow as the source or destination of other data flow, you create data flow chains.
• For database tables, Decision Data Stores, and Adaptive Decision Manager (ADM), use the pxAdaptiveAnalytics data set. For Interaction History (IH) use the pxInteractionHistory data set or Visual Business Director (VBD) data source.
• If your data set type is Decision Data Store (DDS) you can define the period of time for which to store the data. You can also configure the data flow to update only a single property as a result of the data flow run.
• When your data set is a database table, you must specify the save options:
  • Only insert new records
  • Insert new and overwrite existing records

By adding the branch pattern in a data flow, you can perform the following actions:

• Define multiple (secondary) destinations to which you can write data. These destinations can be of various types (for example, different types of data sets).
• Define a destination-specific processing logic.

You insert the branch pattern by adding multiple Destination shapes in a data flow. You can add only one Branch shape in a data flow. The Branch shape radiates connectors that lead to each Destination shape that you created. In each of those connectors, you can add Filter, Convert, and Data Transform shapes to apply processing instructions that are specific only to the destination that the connector leads to.

When you delete the branch pattern, you remove all additional destination patterns and the patterns that are associated with each branch.