xicm opened a new issue, #12015:
URL: https://github.com/apache/iceberg/issues/12015
### Query engine
_No response_
### Question
```
private static class BucketInteger extends Bucket<Integer>
implements SerializableFunction<Integer, Integer> {
private BucketInteger(int numBuckets) {
super(numBuckets);
}
@Override
protected int hash(Integer value) {
return BucketUtil.hash(value);
}
}
```
Why do we use Murmur3 instead of the int value itself?
test code
```
Bucket bucket = Bucket.get(Types.IntegerType.get(), 10);
List<Integer> list = new ArrayList<>();
List<Integer> randomList = new ArrayList<>();
Random rand = new Random();
for (int i = 0; i < 10000000; i++) {
int num = rand.nextInt(10000);
randomList.add(bucket.apply(num));
list.add(bucket.apply(i));
}
System.out.println("natural list");
list.stream()
.collect(Collectors.groupingBy(item -> item,
Collectors.counting()))
.forEach((key, value) -> System.out.println(key + ": " +
value));
System.out.println("random list");
randomList.stream()
.collect(Collectors.groupingBy(item -> item,
Collectors.counting()))
.forEach((key, value) -> System.out.println(key + ": " +
value));
```
result with Murmur3
```
natural list
0: 1000996
1: 999922
2: 1000005
3: 1000139
4: 1001529
5: 998785
6: 999291
7: 999777
8: 999204
9: 1000352
random list
0: 998404
1: 1005602
2: 1021090
3: 1021668
4: 996615
5: 1046078
6: 1023192
7: 997321
8: 923906
9: 966124
```
result of the origin value
```
natural list
0: 1000000
1: 1000000
2: 1000000
3: 1000000
4: 1000000
5: 1000000
6: 1000000
7: 1000000
8: 1000000
9: 1000000
random list
0: 1001911
1: 997982
2: 999024
3: 1000518
4: 999166
5: 999954
6: 1001863
7: 1000033
8: 997981
9: 1001568
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
