When you run a mapping on the Spark engine to read data from a local complex file source, if the folders inside the source directory contains files with same names and you use a wildcard pattern to read the entire parent directory, the Data Integration Service reads data from any one of the files with the same names and there is no impact in reading from the other files.

When you use the FileName port in a complex file target in Binary format and run a mapping in the native environment, the Data Integration Service writes all the files to a single folder.

When you run a mapping to write data to a complex file using

Create target

option for Avro/Parquet formats with mapping flow enabled, the schema is created with primitive datatypes and the Data Integration Service skips the rows having null values.

When you run a mapping on the Spark engine to read data from a complex file with ORC (Optimized Row Columnar) dayatype using the wildcard character questionmark '?' in MapR distro, the mapping fails.

When you validate a mapping and select a connection parameter type that is not valid, the parameter name appears incorrectly.