The model that you select for a data object should match the structure of the expected input files for the data object as much as possible. If an input file does not match the model, or partially matches the model, there might be a large amount of unidentified data and data loss. Therefore, when you create the model, it is important to choose a file that represents the expected input files.

When you create an

intelligent structure model

in

Intelligent Structure Discovery

, do not use duplicate names for different elements.

Ensure that the

intelligent structure model

is valid before you add it to the Column Projection properties of the data object properties.

You can only use a Read transformation with an

intelligent structure model

in a mapping. Do not create or use a Write transformation with an

intelligent structure model

, as the mapping will fail.

When you create the

intelligent structure model

, select the Data Integration Version that corresponds to your current Big Data Management version.

A data object with an

intelligent structure model

parse PDF forms, Microsoft Word tables, and XML files whose size is less than the supported Hadoop split size of 256 MB.

An

intelligent structure model

can parse the data within PDF form fields but not data outside the fields.

An

intelligent structure model

can parse data within Microsoft Word tables. Other data is unidentified.