Github user concretevitamin commented on a diff in the pull request:
https://github.com/apache/spark/pull/1359#discussion_r14782583
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/stringOperations.scala
---
@@ -207,3 +210,64 @@ case class StartsWith(left: Expression, right:
Expression) extends StringCompari
case class EndsWith(left: Expression, right: Expression) extends
StringComparison {
def compare(l: String, r: String) = l.endsWith(r)
}
+
+/**
+ * A function that takes a substring of its first argument starting at a
given position.
+ * Defined for String and Binary types.
+ */
+case class Substring(str: Expression, pos: Expression, len: Expression)
extends Expression {
+
+ type EvaluatedType = Any
+
+ def nullable: Boolean = true
+ def dataType: DataType = {
+ if (str.dataType == BinaryType) str.dataType else StringType
+ }
+
+ def references = children.flatMap(_.references).toSet
+
+ override def children = str :: pos :: len :: Nil
+
+ def slice[T, C <% IndexedSeqOptimized[T,_]](str: C, startPos: Int,
sliceLen: Int): Any = {
+ val len = str.length
+ // Hive and SQL use one-based indexing for SUBSTR arguments but also
accept zero and
--- End diff --
It looks like from [1] and [2] that Hive and SQL by default refer users to
use 1-based indexing (and the links don't seem to mention 0-based at all). Even
though they do support so, this subtle fact necessitates a couple branchings in
the implementation which might cause us some performance penalty. What do you &
people think about supporting 1-based indexing only? +@marmbrus @rxin
---
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.
---
