One more question...
While processing in DataStage do I need some disk space for temporary files or else DataStage uses only memory?
Does it depend on Version of DataStage ( DS Server, DS PX)?
Regards,
ANKa
Disk space / Memory
Moderators: chulett, rschirm, roy
akrzy,
DataStage (both flavors) will use memory and temporary files. Sometimes memory buffers also get temporarily written to disk. It is wise to reserve quite a bit of temporary space for DataStage implementations.
DataStage PX uses more temporary files than does Server, but allows you as a designer to control these more closely.
The temporary space required is highly dependant upon your processing - any sorting or aggregation will use disk space and memory whereas a straightforward job (reading sequential, processing, writing to DB) will use less.
DataStage (both flavors) will use memory and temporary files. Sometimes memory buffers also get temporarily written to disk. It is wise to reserve quite a bit of temporary space for DataStage implementations.
DataStage PX uses more temporary files than does Server, but allows you as a designer to control these more closely.
The temporary space required is highly dependant upon your processing - any sorting or aggregation will use disk space and memory whereas a straightforward job (reading sequential, processing, writing to DB) will use less.
</a>
Code: Select all
DataStage PX uses more temporary files than does Server, but allows you as a designer to control these more closely. 
regards
kumar
Kumar,
yes, the PX total temporary file size is [usually] going to be greater than for a server job. That is the nature of PX - and a reason why each node has it's own temporary file path. Not every job will actually 'land' data in temporary files. If you change your PX job partitioning you are going to need temporary space. If you sort or aggregate you are really going to need space. If you do lookups in PX it will use temp space. Buffered or deferred writes in Server can dump to disk as well.
yes, the PX total temporary file size is [usually] going to be greater than for a server job. That is the nature of PX - and a reason why each node has it's own temporary file path. Not every job will actually 'land' data in temporary files. If you change your PX job partitioning you are going to need temporary space. If you sort or aggregate you are really going to need space. If you do lookups in PX it will use temp space. Buffered or deferred writes in Server can dump to disk as well.