Got it. Seems like something we should get fixed since datagen substantially increases completion time.
Thanks for the clarification. I'll think about potential solutions. -- Jacques Nadeau CTO and Co-Founder, Dremio On Sun, Oct 25, 2015 at 7:11 PM, rahul challapalli < [email protected]> wrote: > Jacques, > > Unfortunately you are not missing anything. There was no drop table support > when these tests were added. So we can only delete them from a script. With > the drop table support in-place you can certainly modify the tests and add > a drop table statement at the end. > > We also have some suites which have CTAS statements in the data gen > scripts. So the script drops and creates the tables required. This is to > avoid running some long running CTAS statements in every single test of the > suite. We at MapR run the data gen step for every single run ( as we still > want to run those CTAS statements). > > - Rahul > > On Sun, Oct 25, 2015 at 4:29 PM, Jacques Nadeau <[email protected]> > wrote: > > > Hey All, > > > > I continue to try to get the test framework up and running. It looks like > > I'm close but I'm seeing something strange. Once I've run the test suite > > once with -d (generate data), I would expect multiple runs to be > successful > > without having to use -d. However, I've noticed with the ctas flatten > tests > > that they are generating new tables that are not ever getting deleted. I > > realized that there is a gen task > > in resources/Functional/ctas/ctas_flatten/100000rows/ctas_flatten.json > that > > is used to the delete tables. This seems to mean that the only way to > > successfully complete a test would be to run the generation step on every > > single run. What am I missing? > > > > thanks, > > Jacques > > > > -- > > Jacques Nadeau > > CTO and Co-Founder, Dremio > > >
