Configure the Projection Node
A Projection node represents a relational selection (filter) combined with a projection operation. It also allows calculated attributes to be added to the output.
- Double-click the Projection node.
Define output columns.
If you have connected the previous node to the Projection node, you can map any columns from the input as output columns of the Projection node. You can also add, delete, and rename the columns, as needed.
In the Mapping pane, select a Source column
and hold and drag the cursor to the Drop row here
zone in the Target section.
- (Optional) If you want to automap the columns based on column names, in the Source section, choose (Auto Map by Name).
If you want to add new output columns, in the Target section,
choose + (Add Column) to add new rows and define
its data type.
If you want to use an expression to define an output column, click the required
In the bottom pane, the modeler opens an expression editor. Use the expression editor to define the output column expression.
If you want to remap an output column with a different source column,
right-click the mapping and choose Remap.
Select a new source column and choose OK.
- If you want to edit a column name or its data type, select a Target column and choose (Edit).
- If you want to delete a mapping without deleting the output column, right-click the mapping and choose Delete.
- In the Mapping pane, select a Source column and hold and drag the cursor to the Drop row here zone in the Target section.
(Optional) Define additional configurations for output columns.
In the Columns tab toolbar, switch to the Form pane to define additional configurations for output columns.
If you have identical records from the previous node, in the toolbar,
select the Distinct checkbox to output unique
- If you want to add new output columns by duplicating an existing output column definition, select an output column and in the toolbar, choose .
Under the Primary Key column, select an output
column that serves as the primary key.
Typically the data in this column is unique.
- Under the Nullable column, choose whether the column value can be empty (nullable).
- To reorder the columns, select the column that you want to move and in the toolbar click the up or down arrows.
- If you have identical records from the previous node, in the toolbar, select the Distinct checkbox to output unique records only.
(Optional) Define filters.
For example, if you want to move all the records that are in Canada for the year 2017, your filter might look like this: "Filter1_input"."COUNTRY" = "Canada" AND "Filter1_input"."DATE"="2017".
- Select the Filters tab to compare the column name against a constant value.
In the expression editor, enter the required expression.
If you want to configure the Data Transform operator with another node,
- In the menu bar, use the breadcrumb navigation to navigate back to the operator configuration editor.
- (Optional) Add new nodes.
- To connect the nodes, select the output port of a node and drag the cursor to an input port of another node.