Column Mapping within Components

Reported for version 10

What is the mapping of output columns within a component?

In DQC, a component is a powerful tool which can be used to perform various subtasks within a plan. There are different ways how to use the column values a component works with. One of the features is column mapping. Through column mapping, the data flow is assigned to the component from a plan. The text fields of a component are used to map the input columns from a plan. They turn into the internal input columns of the component. This applies to the output columns of the component as well. Note, that if a component column is mapped to an input plan column, it is not possible to map the same internal component column as an output to another plan column. The columns from the input to the component are automatically mapped. If working directly with a column is not required, it is not necessary to specify the mapping for those columns inside of the component. DQC will be able to handle the columns properly. If you intend to modify the column values within the component, it is necessary to specify the column in the column mapping.

Data flow within plans is a bit different. When you provide input columns to a component, the values do not need to be returned at all since the component preserves all of the columns. In other words, the data and the columns that are processed in a flow that goes to the component are kept. This does not modify the columns themselves. The data flow in a plan allocates the input columns and allows the component to use them. After the component finishes the work, the input columns are still there. They were used, but there is no need to define them in an output once again. All of the columns that were in the data flow before going to the component will be there after the component finishes processing. 

Within the component, it is also possible to create additional columns. These columns need to be mapped as output columns if they are going to be used after the component finishes the processing. There are some steps which generate auto-mapped values. The auto-mapped columns stay untouched and are accessible after the component finishes processing. The auto-mapping happens because the data format inside the component changes. In simple terms, new records are created from the input data. Columns created in the plan are turned into the output columns. These are formed inside the component because the records before entering the component have different values than after leaving it. Some of the steps with such effect are Join, Representative Creator, Group Aggregator, and Union.

For more information regarding components usage in plans, see the corresponding section in Documentation portal.