This algorithm executes the two given SQL queries on the selected database, returning a graph with the results. The node query should have at least one attribute that is unique for every row in the result. One attribute with that property must be selected as the node id. That attribute can be of any type. The edge query must have two attributes that each have values from the list of node ids. Each source-target pair must be unique (though if the network is directed, you can have both a->b and b->a).
It may be useful to select View and view the database structure.
When giving the node id, source, and target attributes, don't forget that column names from your query will generally be all upper case (unless quoted properly).