Version

    Chapter 57. Transformers

    Common Properties of Transformers

    Transformers are intermediate nodes of the graph.

    Transformers receive data through the connected input port(s), process it in the user-specified way and send it out through the connected output port(s).

    We can distinguish Transformers according to what they can do.

    • One Transformer only copies each input data to all connected outputs.

      • SimpleCopy copies each input data record to all connected output ports.

    • One Transformer passes only some input records to the output.

      • DataSampler passes some input records to the output based on one of the selected filtering strategies.

    • One Transformer removes duplicate data records.

      • Dedup removes duplicate data. Duplicate data can be sent out through the optional second output port.

    • Other components filter data according to the user-defined conditions:

      • Filter compares data with the user-defined condition and sends out records matching this condition. Data records not matching the condition can be sent out through the optional second output port.

    • Other Transformer sort data each in different way:

    • One Transformer is able to aggregate information about data:

      • Aggregate aggregates information about input data records.

    • One Transformer distributes input records among connected output ports:

      • Partition distributes individual input data records among different connected output ports;

      • LoadBalancingPartition distributes incoming input data records among different output ports according workload of downstream components.

    • One Transformer receives data through two input ports and sends it out through three output ports. Data contained in the first port only, in both ports, or in the second port go to corresponding output port.

      • DataIntersection intersects sorted data from two inputs and sends it out through three connected output ports as defined by the intersection.

    • Other Transformers can receive data records from multiple input ports and send them all through the unique output port:

      • Concatenate receives data records with the same metadata from one or more input ports, puts them together and sends them out through the unique output port. Data records from each input port are sent out after all data records from previous input ports;

      • SimpleGather receives data records with the same metadata from one or more input ports, puts them together, and sends them out through the unique output port as fast as possible

      • Merge receives sorted data records with the same metadata from two or more input ports, sorts them all, and sends them out through the unique output port.

    • Other Transformers receive data through connected input port, process it in the user-defined way and send it out through the connected output port(s).

      • Denormalizer creates single output data record from a group of input data records.

      • Pivot is a simple form of Denormalizer which creates a pivot table, summarizing input records.

      • Normalizer creates one or more output data record(s) from a single input data record.

      • MetaPivot works similarly to Normalizer, but it always performs the same transformation and the output metadata is fixed to data types.

      • Reformat processes input data in the user-defined way. Can distribute output data records among different or all connected output ports in the user-defined way.

      • Rollup processes input data in the user-defined way. Can create a number of output records from another number of input records. Can distribute output data records among different or all connected output ports in the user-defined way.

      • DataSampler passes only some input records to the output. You can select from one of the available filtering strategies that suits your needs.

    • One Transformer can transform input data using stylesheets.