Number of input and output link

Number of input link	:	One
Number of output link	:	One or Two

Split Transformator Assembly () functions similarly to Transformator Assembly, but it produces two separate outputs based on the success or failure of the transformation process.

Outputs

Split Transformator Assembly has two output streams:

1. Success Output
   
   • Contains records that are successfully processed by the transformation function
   • The output schema follows the transformation result defined in the assembly
   • These records are forwarded to the normal downstream processing flow
2. Failure Output
   
   • Contains records that fail during transformation
   • Used for error handling, auditing, or reprocessing
   • Failure records include additional information describing why the transformation failed

The following image shows examples of the use of transformation functions in the Split Transformator assembly.

Each input record consists of multiple fields with defined data types. Transformation functions operate on one or more input fields and produce new output fields or records.

Supported Field Data Types

Each field in a record can have one of the following data types:

• String
• Byte
• Short
• Integer
• Float
• Double
• Long
• Timestamp
• BigDecimal
• BigInteger
• Image
• Object (Comparable Object)

Transformation Behavior

• One or more input fields can be transformed into:
  • One output field
  • Multiple output fields
  • Multiple output records
• Transformations are applied record by record
• Output records are passed downstream to the next assembly or sink
• Transformation functions can be chained, as long as they form a straight-line sequence

Transformation Function Groups

Transformation functions are organized into the following groups based on their purpose:

1. Math
2. String
3. Field
4. Lookup
5. Conversion
6. Security
7. Date
8. Text Similarity
9. Genomic
10. Spatial
11. Special Purpose
12. Transpose
13. Content Based Image Retrieval
14. Image Processing
15. Video Processing

Function Chaining Rules

• A transformation function can be connected to another transformation function
• The connection must form a linear (straight-line) flow

Typical Use Cases

• Data cleansing
• Field enrichment
• Parsing and formatting
• Lookup reference
• Pre-processing before Group By

Supported Engines

Map Reduce	Spark 1.x RDD	Spark 2.x RDD	Spark 2.x Dataset	Apache Beam 2.29.0 Batch	Apache Beam 2.29.0 Streaming	Apache Flink 1.12.0 Batch	Apache Flink 1.12.0 Streaming	Apache Tez 0.9.1	PySpark