Github user chenghao-intel commented on a diff in the pull request:
https://github.com/apache/spark/pull/6104#discussion_r30564674
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/WindowFunctionDefinition.scala ---
@@ -0,0 +1,341 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.sql
+
+import scala.language.implicitConversions
+
+import org.apache.spark.annotation.Experimental
+import org.apache.spark.sql.catalyst.expressions._
+
+/**
+ * :: Experimental ::
+ * A set of methods for window function definition for aggregate
expressions.
+ * For example:
+ * {{{
+ * // predefine a window
+ * val w = partitionBy("name").orderBy("id")
+ *
+ * df.select(
+ * first("value")
+ * over(w).as("first_value"),
+ * last("value")
+ * over(w).as("last_value"),
+ * avg("value")
+ * over(
+ * partitionBy("k1")
+ * .orderBy("k2", "k3")
+ * .rows
+ * .following(1)).as("avg_value"),
+ * max("value")
+ * .over(
+ * partitionBy("k2")
+ * .orderBy("k3")
+ * .range
+ * .between
+ * .preceding(4)
+ * .and
+ * .following(3)).as("max_value"))
+ *
+ * }}}
+ *
+ * @param column The bounded the aggregate/window function
+ * @param partitionSpec The partition of the window
+ * @param orderSpec The ordering of the window
+ * @param frame The Window Frame type
+ * @param bindLower A hint of when call the methods `.preceding(n)`
`.currentRow()` `.following()`
+ * if bindLower == true, then we will set the lower
bound, otherwise, we should
+ * set the upper bound for the Row/Range Frame.
+ */
+@Experimental
+class WindowFunctionDefinition protected[sql](
+ column: Column = null,
+ partitionSpec: Seq[Expression] = Nil,
+ orderSpec: Seq[SortOrder] = Nil,
+ frame: WindowFrame = UnspecifiedFrame,
+ bindLower: Boolean = true) {
+
+ private[sql] def newColumn(c: Column): WindowFunctionDefinition = {
+ new WindowFunctionDefinition(c, partitionSpec, orderSpec, frame,
bindLower)
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]] partitioned by the
specified column.
+ * {{{
+ * // The following 2 are equivalent
+ * df.over(partitionBy("k1", "k2", ...))
+ * df.over(partitionBy($"K1", $"k2", ...))
+ * }}}
+ * @group window_funcs
+ */
+ @scala.annotation.varargs
+ def partitionBy(colName: String, colNames: String*):
WindowFunctionDefinition = {
+ partitionBy((colName +: colNames).map(Column(_)): _*)
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]] partitioned by the
specified column. For example:
+ * {{{
+ * df.over(partitionBy($"col1", $"col2"))
+ * }}}
+ * @group window_funcs
+ */
+ @scala.annotation.varargs
+ def partitionBy(cols: Column*): WindowFunctionDefinition = {
+ new WindowFunctionDefinition(column, cols.map(_.expr), orderSpec,
frame)
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]] sorted by the specified
column within
+ * the partition.
+ * {{{
+ * // The following 2 are equivalent
+ * df.over(partitionBy("k1").orderBy("k2", "k3"))
+ * df.over(partitionBy("k1").orderBy($"k2", $"k3"))
+ * }}}
+ * @group window_funcs
+ */
+ @scala.annotation.varargs
+ def orderBy(colName: String, colNames: String*):
WindowFunctionDefinition = {
+ orderBy((colName +: colNames).map(Column(_)): _*)
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]] sorted by the specified
column within
+ * the partition. For example
+ * {{{
+ * df.over(partitionBy("k1").orderBy($"k2", $"k3"))
+ * }}}
+ * @group window_funcs
+ */
+ def orderBy(cols: Column*): WindowFunctionDefinition = {
+ val sortOrder: Seq[SortOrder] = cols.map { col =>
+ col.expr match {
+ case expr: SortOrder =>
+ expr
+ case expr: Expression =>
+ SortOrder(expr, Ascending)
+ }
+ }
+ new WindowFunctionDefinition(column, partitionSpec, sortOrder, frame)
+ }
+
+ /**
+ * Returns the current [[WindowFunctionDefinition]]. This is a dummy
function,
+ * which makes the usage more like the SQL.
+ * For example:
+ * {{{
+ * df.over(partitionBy("k1").orderBy($"k2",
$"k3").range.between.preceding(1).and.currentRow)
+ * }}}
+ * @group window_funcs
+ */
+ def between: WindowFunctionDefinition = {
+ assert(this.frame.isInstanceOf[SpecifiedWindowFrame], "Should be a
WindowFrame.")
+ new WindowFunctionDefinition(column, partitionSpec, orderSpec, frame,
true)
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]] indicate that we need to
specify the
+ * upper bound.
+ * For example:
+ * {{{
+ * df.over(partitionBy("k1").orderBy($"k2",
$"k3").range.between.preceding(3).and.currentRow)
+ * }}}
+ * @group window_funcs
+ */
+ def and: WindowFunctionDefinition = {
+ new WindowFunctionDefinition(column, partitionSpec, orderSpec, frame,
false)
+ }
+
+ /**
+ * Returns a new Ranged [[WindowFunctionDefinition]].
+ * For example:
+ * {{{
+ * df.over(partitionBy("k1").orderBy($"k2",
$"k3").range.between.preceding(3).and.currentRow)
+ * }}}
+ * @group window_funcs
+ */
+ def range: WindowFunctionDefinition = {
+ new WindowFunctionDefinition(column, partitionSpec, orderSpec,
+ SpecifiedWindowFrame(RangeFrame, UnboundedPreceding,
UnboundedFollowing))
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]], with fixed number of
records.
+ * For example:
+ * {{{
+ * df.over(partitionBy("k1").orderBy($"k2", $"k3").rows)
+ * }}}
+ * @group window_funcs
+ */
+ def rows: WindowFunctionDefinition = {
+ new WindowFunctionDefinition(column, partitionSpec, orderSpec,
+ SpecifiedWindowFrame(RowFrame, UnboundedPreceding,
UnboundedFollowing))
+ }
+
+ /**
+ * Returns a new [[WindowFunctionDefinition]], with position specified
preceding of CURRENT_ROW.
+ * It can be either Lower or Upper Bound position depends on the
semantic context.
+ * For example:
+ * {{{
+ * // [CURRENT_ROW - 1, ~)
+ * df.over(partitionBy("k1").orderBy("k2").rows.preceding(1))
+ * // [CURRENT_ROW - 3, CURRENT_ROW - 1]
+ *
df.over(partitionBy("k1").orderBy("k2").rows.between.preceding(3).and.preceding(1))
+ * // (~, CURRENT_ROW - 1]
+ *
df.over(partitionBy("k1").orderBy("k2").rows.between.unboundedPreceding.and.preceding(1))
+ * }}}
+ * @group window_funcs
+ */
+ def preceding(n: Int): WindowFunctionDefinition = {
+ assert(n > 0)
--- End diff --
Just confirm, Hive will take `0 preceding` as `CURRENT_ROW`, I will follow
the same pattern
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
