Github user nongli commented on a diff in the pull request:
https://github.com/apache/spark/pull/10435#discussion_r48778819
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/misc.scala
---
@@ -177,3 +179,44 @@ case class Crc32(child: Expression) extends
UnaryExpression with ImplicitCastInp
})
}
}
+
+/**
+ * A function that calculates hash value for a group of expressions.
+ *
+ * Internally this function will write arguments into an [[UnsafeRow]],
and calculate hash code of
+ * the unsafe row using murmur3 hasher with a seed.
+ * We should use this hash function for both shuffle and bucket, so that
we can guarantee shuffle
+ * and bucketing have same data distribution.
+ */
+case class Murmur3Hash(children: Seq[Expression], seed: Int) extends
Expression {
+ def this(arguments: Seq[Expression]) = this(arguments, 42)
--- End diff --
I think this is fine.
Can you file a follow up jira to look at this again? I think we want to
remove the projection to unsafe row soon (before we ship this and persist
metadata that way). This should be decoupled from unsafe row ideally. For
example, if the row is (int, double, string): the generated hash function
shoudl be something like
int hash = seed;
hash = murmur3(getInt(0), hash)
hash = murmur3(getDouble(1), hash)
hash = murmur3(getString(2), hash)
return hash
This is likely not the currently computed hash value so can't defer this
for too long.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
