A Linker

provides linking between two input tables. Typically, the two input tables are heterogenous and share a key column. Most of the time the keys are primary, and each of the two tables has exactly one data item with a unique key. But it is also possible to use keys that do not have one-to-one mapping. For example, to find incident edges of a node in a graph, the node ID may be used as a key, which corresponds to multiple edges in the edge table.

The table T1 that connects to the first Input Port provides the shared keys. These keys are extracted from the Extract Column of T1.

The table T2 that connects to the second Input Port gives the subset to be filtered. The data items in T2 are kept when they have a Filter Column value that equals some value in the extracted keys from T1.

A linker is equivalent to a combination of one Constants Generator

and one Attribute Filter
, where the constants generator extracts the keys from T1 that are used to filter T2. Using a linker may simply diagram connections.


Extract Column

Configures the column from which the keys are extracted. The column is from the first table T1 that is connected to the first Input Port .

Filter Column

Configures the column on which filtering is applied on the second table T2 that is connected to the second Input Port . The filtering matches the attribute values in this column against the keys extracted from T1 using Full String matching.