loading utf-8 messages via jdbc connectivity
when messages containing the � replacement char for untranslatable character
with our DBS Control param 104 (AcceptReplacementCharacters) set to false
jdbc returns error 1338, 1339.
turning DBS Control param 104 (AcceptReplacementCharacters) to true would disable partition elimination for character based partition.
I do not really understand the reason provided by Paul Sinclair -
http://developer.teradata.com/blog/paulsinclair/2012/07/td-14-0-the-other-partitioning-enhancements - might be due to my limited knowledge of things
At the moment my perception is:
since once the replacement character is accepted into the Teradata environment
its value remains to be evaluated a 'replacement character' for ever independently of whether the set of untranslatable characterS for a given character set changes over the time.
this should then no be a consideration to cut the partition elimination of character based partition.
Other point of views / corrections expected
rgds - JL D
Morning thought :
is the argument still valid in the case
you consider replacing the replacement character with a dot . for instance in order to get the data in Teradata
You would end up with the same situation – from a analysis perspective – AND a working a “non eliminated” char partition .
but you would have had to parse the data stream once more reducing performance in order to replace the replacing character.
rgds - JL D