Re: Hi, Hive People urgent question about [Distribute By] function

Hello, the same question about DISTRIBUTE BY on Hive. Accorring to you, you do not use hashCode of Object class on DBY, Distribute By. I tried to understand how ObjectInspectorUtils works for distribution, but it seemed it has a lot of Hive API. It is not much understnading. I want to override

Re: Hi, Hive People urgent question about [Distribute By] function

> I want to override partitionByHash function on Flink like the same way >of DBY on Hive. > I am working on implementing some benchmark system for these two system, >which could be contritbutino to Hive as well. I would be very disappointed if Flink fails to outperform Hive with a Distribute BY,

Re: Hi, Hive People urgent question about [Distribute By] function

Hello, the same question about DISTRIBUTE BY on Hive. Accorring to you, you do not use hashCode of Object class on DBY, Distribute By. I tried to understand how ObjectInspectorUtils works for distribution, but it seemed it has a lot of Hive API. It is not much understnading. I want to override

Re: Hi, Hive People urgent question about [Distribute By] function

> so do you think if we want the same result from Hive and Spark or the >other freamwork, how could we try this one ? There's a special backwards compat slow codepath that gets triggered if you do set mapred.reduce.tasks=199; (or any number) This will produce the exact same hash-code as the

Hi, Hive People urgent question about [Distribute By] function

Hello, I am working on Flink and Spark majoring in Computer Science in Berlin. I have the important question. Well, this question is from what I do these days, which is translations Hive Query to Flink. When applying [Distribute By] on Hive to the framework, the function should be

Re: Hi, Hive People urgent question about [Distribute By] function

Thanks for your help. so do you think if we want the same result from Hive and Spark or the other freamwork, how could we try this one ? could you tell me in detail. Regards, Philip On Thu, Oct 22, 2015 at 6:25 PM, Gopal Vijayaraghavan wrote: > > > When applying

Re: Hi, Hive People urgent question about [Distribute By] function

> When applying [Distribute By] on Hive to the framework, the function >should be partitionByHash on Flink. This is to spread out all the rows >distributed by a hash key from Object Class in Java. Hive does not use the Object hashCode - the identityHashCode is inconsistent, so Object.hashCode()