Hi,


Just want to make you aware of the following JIRA:
https://issues.apache.org/jira/browse/TRAFODION-1775.



In that JIRA, I propose making it an error to do UPDATE STATISTICS on a
really big table without using sampling. The reason being that it can take
a very long time in this case and is usually not intended. I will provide a
syntax that can override the error.



If you think this is a bad idea, or you think of better approaches, please
comment in the JIRA. If you think it is a great idea, you can make that
comment too, but I will consider silence as consent.



Thanks,



Dave

Reply via email to