hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexander Pivovarov" <apivova...@gmail.com>
Subject Re: Review Request 34586: HIVE-10704
Date Sat, 30 May 2015 02:57:36 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/34586/#review85853
-----------------------------------------------------------



ql/src/java/org/apache/hadoop/hive/ql/exec/tez/HashTableLoader.java
<https://reviews.apache.org/r/34586/#comment137688>

    The comment above says - "if any table has bad size estimate...".
    But why you check "totalSize <= 0" then?
    Should you iterate over all small tables and check that they all have good size estimate.
    
    What if you have table sizes (100, -4, 0)
    totalSize is 96. But table #2 size is -4, which is bad size.
    
    To make code clear I recommend to add new boolean variable isAnyTableHasBadSize and set
its value it in the place where you calc totalSize, biggest and maxSize


- Alexander Pivovarov


On May 27, 2015, 6:33 a.m., Mostafa Mokhtar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/34586/
> -----------------------------------------------------------
> 
> (Updated May 27, 2015, 6:33 a.m.)
> 
> 
> Review request for hive.
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> fix biggest small table selection when table sizes are 0
> fallback to dividing memory equally if any tables have invalid size
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/tez/HashTableLoader.java 536b92c 
> 
> Diff: https://reviews.apache.org/r/34586/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Mostafa Mokhtar
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message