Hiļ¼all
Recently i want to test carbon1.1.1+spark2.1.0 performance and test
environment already have carbon1.1.0+spark1.6 and carbon1.1.1+spark1.6 and
carbon tables generated.
I tried reading carbon tables generated from carbon1.1.0+spark1.6 and
carbon1.1.1+spark1.6 as well as delete in subquery data generated from
carbon1.1.1+spark1.6 using carbon1.1.1+spark2.1.0 roughly and all worked ok.
My question is do they have compatibility problems and may impare
performance?Or should I regenerate carbon table using carbon1.1.1+spark2.1.0 ?
And I break my question into 3 parts:
1.Will IUD and other opeartion(like sub-query) works fine if use
carbon1.1.1+spark2.1.0 to operate carbon1.1.0+spark1.6?
2.Will use carbon1.1.1+spark2.1.0 to operate carbon1.1.0+spark1.6
impare performance or should regenerate data using carbon1.1.1+spark2.1.0
while tesing carbon1.1.1+spark2.1.0?
3.Will use carbon1.1.1+spark2.1.0 to operate carbon1.1.1+spark1.6
impare performance or should regenerate data using carbon1.1.1+spark2.1.0
while tesing carbon1.1.1+spark2.1.0?
[email protected]
