If you read the help document, you will come to know that the sequential file stage in a server job will support multiple input and output links. And you can have one file for each of the output link. Use a link collector to get the data into one link.
This is use full only if the number of files remain constant.
In parallel job Sequential File stage you can have multiple File properties for reading specific files, for File Pattern for reading files whose names match a given regular expression.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.