We assume that you already know what components are. See Chapter 26, Components for brief information.
Some of the components are intermediate nodes of the graph. These are called Transformers or Joiners.
For information about Joiners see Chapter 56, Joiners. Here we will describe Transformers.
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).
Components can have different properties. But they also can have some in common. Some properties are common for all of them, others are common for most of the components, or they are common for Transformers only. You should learn:
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:
ExtFilter 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.
EmailFilter validates e-mail adresses and sends out the valid ones. Data records with invalid e-mail addresses can be sent out through the optional second output port.
Other Transformer sort data each in different way:
ExtSort sorts input data.
FastSort sorts input data faster than ExtSort.
SortWithinGroups sorts input data withing groups of sorted data.
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.
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.
XSLTransformer transforms input data using stylesheets.