Skip to main content

Prepare Recommendations

Recommendations and Scripts provide rich technical details and tips about the components, usage, and performance of prepare dataflows.

The dataflow must be saved before Recommendations can be and accessed. Even if only one entity has been added to the canvas, the dataflow must be saved to access helpful design tips. For example, when JDBC entities are discovered at the same time, relations between fields are identified and retained for Primary Keys (PK) and Foreign Keys (FK). This information is leveraged in the metadata to assist users to build dataflows.

RECOMMENDATIONS: COMMON USAGE tab details how the same entities have been used in other dataflows.

RECOMMENDATIONS: DESIGN TIPS example for a JDBC entity that was discovered with other entities: Primary and Foreign keys provide direction for Joins.Suggestions are provided for building workflows like Joiner conditions, Filter and Router options, Transformer Ranges and Trimming options.

RECOMMENDATIONS: PERFORMANCE TIPS tab details how users can optimize performance for dataflow execution

Note: Performance recommendations are generated when the latest execution of a dataflow is in Mapreduce mode. (Mapreduce mode is when Mapreduce is invoked in the back-end to perform a particular operation on the data that exists in the file system). However, if the dataflow is simple, no performance recommendations may be generated.