Github user cloud-fan commented on a diff in the pull request:
https://github.com/apache/spark/pull/19222#discussion_r175592083
--- Diff:
common/unsafe/src/main/java/org/apache/spark/unsafe/hash/Murmur3_x86_32.java ---
@@ -49,55 +51,81 @@ public static int hashInt(int input, int seed) {
}
public int hashUnsafeWords(Object base, long offset, int lengthInBytes) {
- return hashUnsafeWords(base, offset, lengthInBytes, seed);
+ return hashUnsafeWordsBlock(MemoryBlock.allocateFromObject(base,
offset, lengthInBytes), seed);
+ }
+
+ public static int hashUnsafeWordsBlock(MemoryBlock base, int seed) {
+ // This is based on Guava's
`Murmur32_Hasher.processRemaining(ByteBuffer)` method.
+ int lengthInBytes = Ints.checkedCast(base.size());
+ assert (lengthInBytes % 8 == 0): "lengthInBytes must be a multiple of
8 (word-aligned)";
+ int h1 = hashBytesByIntBlock(base, seed);
+ return fmix(h1, lengthInBytes);
}
public static int hashUnsafeWords(Object base, long offset, int
lengthInBytes, int seed) {
// This is based on Guava's
`Murmur32_Hasher.processRemaining(ByteBuffer)` method.
assert (lengthInBytes % 8 == 0): "lengthInBytes must be a multiple of
8 (word-aligned)";
- int h1 = hashBytesByInt(base, offset, lengthInBytes, seed);
+ int h1 = hashBytesByIntBlock(MemoryBlock.allocateFromObject(base,
offset, lengthInBytes), seed);
--- End diff --
It's more consistent to call `hashUnsafeWordsBlock` here.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
