Fan Trap
A measure is a numeric field that can be aggregated. Aggregation of fields at the detail level is always acceptable, but incorrect results can be produced when a summary-level fields are aggregated. The problem of these higher-level aggregates is known as the fan trap.
For example, consider an “Order” table containing total order amounts. If this table is joined with an “Order Items” table, then aggregation by order item will recount the order amount multiple times if an order contains more than one item.

This trap can be readily identified by relationship cardinality (indicated by the “fan” symbol on the join terminus). If the specified measure exists in the “1” side table that has a 1:n relationship with another table, this trap will occur.
The fan trap is evidence of an inherent deficiency in this particular schema, wherein the schema is not providing the order amount on the item-level of granularity. If this aggregation is desired, the best option is to enhance the schema so that the order amount is broken down and recorded in the 'Order Items' table.
If this aggregation is not desired, use two physical views to separate the measure and the lower-granularity table to prevent incorrect aggregation. You can still create a data view that shows both order price and all associated items by building the desired Data Block within a Data Mashup.
| << Loop Traps | © 1996-2013 InetSoft Technology Corporation (v11.5) | Chasm Trap (Cardinality Trap) >> |