From 7ea6d282b925819ddb3874a67b3c9da8cc41f131 Mon Sep 17 00:00:00 2001 From: Cheng Lian Date: Mon, 25 Jul 2016 09:42:39 -0700 Subject: [SPARK-16703][SQL] Remove extra whitespace in SQL generation for window functions ## What changes were proposed in this pull request? This PR fixes a minor formatting issue of `WindowSpecDefinition.sql` when no partitioning expressions are present. Before: ```sql ( ORDER BY `a` ASC ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) ``` After: ```sql (ORDER BY `a` ASC ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) ``` ## How was this patch tested? New test case added in `ExpressionSQLBuilderSuite`. Author: Cheng Lian Closes #14334 from liancheng/window-spec-sql-format. --- .../sqlgen/aggregate_functions_and_window.sql | 2 +- .../sqlgen/regular_expressions_and_window.sql | 2 +- .../src/test/resources/sqlgen/window_basic_1.sql | 2 +- .../src/test/resources/sqlgen/window_basic_2.sql | 2 +- .../sql/catalyst/ExpressionSQLBuilderSuite.scala | 35 ++++++++++++++++++++-- 5 files changed, 37 insertions(+), 6 deletions(-) (limited to 'sql/hive') diff --git a/sql/hive/src/test/resources/sqlgen/aggregate_functions_and_window.sql b/sql/hive/src/test/resources/sqlgen/aggregate_functions_and_window.sql index 3a29bcf2ae..c94f53b0a9 100644 --- a/sql/hive/src/test/resources/sqlgen/aggregate_functions_and_window.sql +++ b/sql/hive/src/test/resources/sqlgen/aggregate_functions_and_window.sql @@ -1,4 +1,4 @@ -- This file is automatically generated by LogicalPlanToSQLSuite. SELECT MAX(c) + COUNT(a) OVER () FROM parquet_t2 GROUP BY a, b -------------------------------------------------------------------------------- -SELECT `gen_attr` AS `(max(c) + count(a) OVER ( ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING))` FROM (SELECT (`gen_attr` + `gen_attr`) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, count(`gen_attr`) OVER ( ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT max(`gen_attr`) AS `gen_attr`, `gen_attr` FROM (SELECT `a` AS `gen_attr`, `b` AS `gen_attr`, `c` AS `gen_attr`, `d` AS `gen_attr` FROM `default`.`parquet_t2`) AS gen_subquery_0 GROUP BY `gen_attr`, `gen_attr`) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 +SELECT `gen_attr` AS `(max(c) + count(a) OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING))` FROM (SELECT (`gen_attr` + `gen_attr`) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, count(`gen_attr`) OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT max(`gen_attr`) AS `gen_attr`, `gen_attr` FROM (SELECT `a` AS `gen_attr`, `b` AS `gen_attr`, `c` AS `gen_attr`, `d` AS `gen_attr` FROM `default`.`parquet_t2`) AS gen_subquery_0 GROUP BY `gen_attr`, `gen_attr`) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 diff --git a/sql/hive/src/test/resources/sqlgen/regular_expressions_and_window.sql b/sql/hive/src/test/resources/sqlgen/regular_expressions_and_window.sql index 7d837d490f..a7909485fb 100644 --- a/sql/hive/src/test/resources/sqlgen/regular_expressions_and_window.sql +++ b/sql/hive/src/test/resources/sqlgen/regular_expressions_and_window.sql @@ -1,4 +1,4 @@ -- This file is automatically generated by LogicalPlanToSQLSuite. SELECT MAX(key) OVER (PARTITION BY key % 3) + key FROM parquet_t1 -------------------------------------------------------------------------------- -SELECT `gen_attr` AS `(max(key) OVER (PARTITION BY (key % CAST(3 AS BIGINT)) ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) + key)` FROM (SELECT (`gen_attr` + `gen_attr`) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, max(`gen_attr`) OVER (PARTITION BY `gen_attr` ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, (`gen_attr` % CAST(3 AS BIGINT)) AS `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 +SELECT `gen_attr` AS `(max(key) OVER (PARTITION BY (key % CAST(3 AS BIGINT)) ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) + key)` FROM (SELECT (`gen_attr` + `gen_attr`) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, max(`gen_attr`) OVER (PARTITION BY `gen_attr` ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, (`gen_attr` % CAST(3 AS BIGINT)) AS `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 diff --git a/sql/hive/src/test/resources/sqlgen/window_basic_1.sql b/sql/hive/src/test/resources/sqlgen/window_basic_1.sql index 73f343d101..f86ca5e086 100644 --- a/sql/hive/src/test/resources/sqlgen/window_basic_1.sql +++ b/sql/hive/src/test/resources/sqlgen/window_basic_1.sql @@ -1,4 +1,4 @@ -- This file is automatically generated by LogicalPlanToSQLSuite. SELECT MAX(value) OVER (PARTITION BY key % 3) FROM parquet_t1 -------------------------------------------------------------------------------- -SELECT `gen_attr` AS `max(value) OVER (PARTITION BY (key % CAST(3 AS BIGINT)) ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING)` FROM (SELECT `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, max(`gen_attr`) OVER (PARTITION BY `gen_attr` ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, (`gen_attr` % CAST(3 AS BIGINT)) AS `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 +SELECT `gen_attr` AS `max(value) OVER (PARTITION BY (key % CAST(3 AS BIGINT)) ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING)` FROM (SELECT `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, max(`gen_attr`) OVER (PARTITION BY `gen_attr` ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, (`gen_attr` % CAST(3 AS BIGINT)) AS `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2) AS gen_subquery_3 diff --git a/sql/hive/src/test/resources/sqlgen/window_basic_2.sql b/sql/hive/src/test/resources/sqlgen/window_basic_2.sql index 3ddb87f91b..15559ac8e1 100644 --- a/sql/hive/src/test/resources/sqlgen/window_basic_2.sql +++ b/sql/hive/src/test/resources/sqlgen/window_basic_2.sql @@ -2,4 +2,4 @@ SELECT key, value, ROUND(AVG(key) OVER (), 2) FROM parquet_t1 ORDER BY key -------------------------------------------------------------------------------- -SELECT `gen_attr` AS `key`, `gen_attr` AS `value`, `gen_attr` AS `round(avg(key) OVER ( ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING), 2)` FROM (SELECT `gen_attr`, `gen_attr`, round(`gen_attr`, 2) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, avg(`gen_attr`) OVER ( ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2 ORDER BY `gen_attr` ASC) AS parquet_t1 +SELECT `gen_attr` AS `key`, `gen_attr` AS `value`, `gen_attr` AS `round(avg(key) OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING), 2)` FROM (SELECT `gen_attr`, `gen_attr`, round(`gen_attr`, 2) AS `gen_attr` FROM (SELECT gen_subquery_1.`gen_attr`, gen_subquery_1.`gen_attr`, avg(`gen_attr`) OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS `gen_attr` FROM (SELECT `gen_attr`, `gen_attr` FROM (SELECT `key` AS `gen_attr`, `value` AS `gen_attr` FROM `default`.`parquet_t1`) AS gen_subquery_0) AS gen_subquery_1) AS gen_subquery_2 ORDER BY `gen_attr` ASC) AS parquet_t1 diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/ExpressionSQLBuilderSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/ExpressionSQLBuilderSuite.scala index c8bf20d13b..fef726c5d8 100644 --- a/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/ExpressionSQLBuilderSuite.scala +++ b/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/ExpressionSQLBuilderSuite.scala @@ -20,8 +20,7 @@ package org.apache.spark.sql.catalyst import java.sql.Timestamp import org.apache.spark.sql.catalyst.dsl.expressions._ -import org.apache.spark.sql.catalyst.expressions.{If, Literal} - +import org.apache.spark.sql.catalyst.expressions.{If, Literal, SpecifiedWindowFrame, WindowSpecDefinition} class ExpressionSQLBuilderSuite extends SQLBuilderTest { test("literal") { @@ -79,4 +78,36 @@ class ExpressionSQLBuilderSuite extends SQLBuilderTest { checkSQL(-'a.int, "(-`a`)") checkSQL(-('a.int + 'b.int), "(-(`a` + `b`))") } + + test("window specification") { + val frame = SpecifiedWindowFrame.defaultWindowFrame( + hasOrderSpecification = true, + acceptWindowFrame = true + ) + + checkSQL( + WindowSpecDefinition('a.int :: Nil, Nil, frame), + s"(PARTITION BY `a` $frame)" + ) + + checkSQL( + WindowSpecDefinition('a.int :: 'b.string :: Nil, Nil, frame), + s"(PARTITION BY `a`, `b` $frame)" + ) + + checkSQL( + WindowSpecDefinition(Nil, 'a.int.asc :: Nil, frame), + s"(ORDER BY `a` ASC $frame)" + ) + + checkSQL( + WindowSpecDefinition(Nil, 'a.int.asc :: 'b.string.desc :: Nil, frame), + s"(ORDER BY `a` ASC, `b` DESC $frame)" + ) + + checkSQL( + WindowSpecDefinition('a.int :: 'b.string :: Nil, 'c.int.asc :: 'd.string.desc :: Nil, frame), + s"(PARTITION BY `a`, `b` ORDER BY `c` ASC, `d` DESC $frame)" + ) + } } -- cgit v1.2.3