After you provide an input file,

Intelligent Structure Discovery

determines the underlying and repeating patterns of the data and creates a structure that represents the fields of data and their relationships. You can quickly model data for files whose structure is very hard, time consuming, and costly to find, such as log files, clickstreams, customer web access, error text files, or other internet, sensor, or device data that does not follow industry standards.

The following image shows the process by which

Intelligent Structure Discovery

deciphers the underlying patterns of data and creates a model of the data patterns:

When the model is based on input from a Hadoop Files source,

Intelligent Structure Discovery

determines the data types to use in the model based on the types and structures in the Hadoop schema. In the model,

Intelligent Structure Discovery

creates an array of structures of key-value pairs for data of type map. Nodes of type enum are represented as strings.