This question was asked by a business partner. I let him know that I would post it to the Forum for response by Streams Development and from other customers/partners.
We are using IBM InfoSphere streams virtual machine. we need technical help to understand IBM streams.
We are trying to ingest some records in IBM streams through 'netcat' utility, here we are using 'TCPSource' to collect the incoming records.
We are doing aggregation on records. And output it in' FileSink'.
Following are the difficulties we are facing in IBM InfoSphere Streams
1. How can we handle error records in TCP source
2. TCP source should not get closed even if the erroneous record comes
3. Empty lines in incoming record should be considered as a tuple with NULL/zero values rather than neglecting it
4. If no input comes in particular window time then it should give zero/NULL as a output
Please let us know the possible solutions.