Apache
/
druid
mirror of https://github.com/apache/druid.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

GroupBy v2: Improve hash code distribution. (#3407) 

Without this transformation, distribution of hash % X is poor in general.
It is catastrophically poor when X is a multiple of 31 (many slots would
be empty).
This commit is contained in:
Gian Merlino 2016-08-29 23:39:08 -07:00 committed by Nishant
parent f037dfcaa4
commit b11e9544ea
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
processing/src/main/java/io/druid/query/groupby/epinephelinae/Groupers.java
View File

@ -46,7 +46,9 @@ public class Groupers
public static int hash(final Object obj)
{
// Mask off the high bit so we can use that to determine if a bucket is used or not.
return obj.hashCode() & 0x7fffffff;
// Also apply the same XOR transformation that j.u.HashMap applies, to improve distribution.
final int code = obj.hashCode();
return (code ^ (code >>> 16)) & 0x7fffffff;
}
public static <KeyType extends Comparable<KeyType>> Iterator<Grouper.Entry<KeyType>> mergeIterators(