Input step

The Input step describes how the links and columns of a relational data structure are represented as a hierarchical data structure.

The Input step provides two views: a Links view and a Tree view.

The Links view displays the same column definition table that you see in on the Columns tab of a stage. From the Links view, you can edit the column definitions and then view the result in the step Output. When you save the assembly, the edited column definitions are saved in the job properties.

The Tree view displays the relational links structure after it is transformed into a hierarchical structure. The root element is Top, and it contains the entire input data. Each input link is transformed into a child list item of the InputLinks node. Each column is transformed into a content item and assigned a primitive type. For example, a column that is assigned the SQL Varchar type is transformed into a content item that has the String data type.

The step Output of the Input step becomes the step Input for the next step in the assembly. This process of taking the transformed data structure from the previous step; transforming, parsing, or composing data; and then producing a modified data structure for the next step continues through each step in the assembly. At each step, you can look at the step Input, which displays the result of the previous step, and the step Output, which displays the output results of the current step, to assess the changes that each step makes to the data structure.

Not all assemblies use an Input step. For example, if the Hierarchical Data stage is used as a source stage in a job, there are no input links and the Input step is empty. Even if the Input step is empty, it remains a permanent part of an assembly; you cannot remove it. Parallel jobs can receive input from multiple input links, while server jobs can receive input from only a single input link.