This section describes the changes to transformations in Big Data Streaming in version 10.2.2 Service Pack 1.
Rank Transformation
Effective in version 10.2.2 Service Pack 1, a streaming mapping must meet the following additional requirements if it contains a Rank transformation:
A streaming mapping cannot contain a Rank transformation and a passive Lookup transformation that is configured with an inequality lookup condition in the same pipeline. Previously, you could use a Rank transformation and a passive Lookup transformation that is configured with an inequality lookup condition in the same pipeline.
A Rank transformation in a streaming mapping cannot have a downstream Joiner transformation. Previously, you could use a Rank transformation anywhere before a Joiner transformation in a streaming mapping.
A streaming mapping cannot contain more than one Rank transformation in the same pipeline. Previously, you could use multiple Rank transformations in a streaming mapping.
A streaming mapping cannot contain an Aggregator transformation and a Rank transformation in the same pipeline. Previously, you could use an Aggregator transformation and a Rank transformation in the same pipeline.
Sorter Transformation
Effective in version 10.2.2 Service Pack 1, a streaming mapping must meet the following additional requirements if it contains a Sorter transformation:
A streaming mapping runs in complete output mode if it contains a Sorter transformation. Previously, a streaming mapping used to run in append output mode if it contains a Sorter transformation.
The Sorter transformation in a streaming mapping must have an upstream Aggregator transformation. Previously, you could use a Sorter transformation without an upstream Aggregator transformation.
The Window transformation upstream from an Aggregator transformation will be ignored if the mapping contains a Sorter transformation. Previously, the Window transformation upstream from an Aggregator transformation was not ignored if the mapping contains a Sorter transformation.
