I'm experimenting with PARALLEL(=6 in my case) and ...%u.dmp to create faster datapump export files with expdp.
So far, I've seen a great improvement. And I know that, for impdp, it seems one must specify DUMPFILE=file1.dmp,file2.dmp, etc.
However. Would the log file and/or any internal dump file metadata, provide sufficient info to allow importing a specific single table *without* referencing all 6 dump files using impdp on the command line. Is this even a valid question to ask? In other words, is there any value in trying to identify a single constituent file for the dump of an entire schema (in my use case) when importing a single table? I'm not sure of all the details, but was asked about this and didn't know how to answer.
My suspicion is that you would, and should, specify all constituent dump files if you've got a series of files resulting from a single expdp command, using the PARALLEL switch with ...%u.dmp, when importing. But I could be wrong and often am. Thanks very much.