[jira] [Comment Edited] (HIVE-19501) Fix HyperLogLog to be threadsafe

["Gopal V (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-19501?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16472950#comment-16472950 ] 

Gopal V edited comment on HIVE-19501 at 5/12/18 7:10 AM:
---------------------------------------------------------

The short and int versions are nearly identical, except they need to account for the sign-bit during the cast up from int -> long by doing 

{code}
  public static long hash64(int data) {
    long k1 = Integer.reverseBytes(data) & (-1L >>> 32);
    int length = Integer.BYTES;
    long hash = DEFAULT_SEED;
    k1 *= C1;
    k1 = Long.rotateLeft(k1, R1);
    k1 *= C2;
    hash ^= k1;
    // finalization
    hash ^= length;
    hash = fmix64(hash);
    return hash;
  }
{code}



was (Author: gopalv):
The short and int versions are nearly identical, except they need to account for the sign-bit during the cast up from int -> long by doing 

{code}
long k = Integer.reverseBytes(data) & (-1L >>> 32);
int length = Integer.BYTES;
{code}


> Fix HyperLogLog to be threadsafe
> --------------------------------
>
>                 Key: HIVE-19501
>                 URL: https://issues.apache.org/jira/browse/HIVE-19501
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Zoltan Haindrich
>            Priority: Major
>
> not sure if this is an issue in reality or not; but there are 3 static fields in HyperLogLog which are rewritten during working; if there are multiple threads are calculating HLL in the same JVM, there is a theoretical chance that they might overwrite eachothers value...
> static fields:
> https://github.com/apache/hive/blob/8028ce8a4cf5a03e2998c33e032a511fae770b47/standalone-metastore/src/main/java/org/apache/hadoop/hive/common/ndv/hll/HyperLogLog.java#L65
> usage:
> https://github.com/apache/hive/blob/8028ce8a4cf5a03e2998c33e032a511fae770b47/standalone-metastore/src/main/java/org/apache/hadoop/hive/common/ndv/hll/HyperLogLog.java#L216



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)