Make the behavior of Postgre dialect independent of ansi mode config

xuanyuanking · xuanyuanking · commit a25088680619 · 2019-11-19T04:30:25.000+08:00
diff --git a/sql/catalyst/src/main/antlr4/org/apache/spark/sql/catalyst/parser/SqlBase.g4 b/sql/catalyst/src/main/antlr4/org/apache/spark/sql/catalyst/parser/SqlBase.g4
@@ -748,7 +748,7 @@ primaryExpression
     | qualifiedName '.' ASTERISK                                                               #star
     | '(' namedExpression (',' namedExpression)+ ')'                                           #rowConstructor
     | '(' query ')'                                                                            #subqueryExpression
-    | qualifiedName '(' (setQuantifier? argument+=expression (',' argument+=expression)*)? ')'
+    | functionCallName '(' (setQuantifier? argument+=expression (',' argument+=expression)*)? ')'
        (OVER windowSpec)?                                                                      #functionCall
     | identifier '->' expression                                                               #lambda
     | '(' identifier (',' identifier)+ ')' '->' expression                                     #lambda
@@ -912,6 +912,12 @@ qualifiedNameList
     : qualifiedName (',' qualifiedName)*
     ;
 
+functionCallName
+    : qualifiedName
+    | LEFT
+    | RIGHT
+    ;
+
 qualifiedName
     : identifier ('.' identifier)*
     ;
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/AstBuilder.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/AstBuilder.scala
@@ -1585,7 +1585,7 @@ class AstBuilder(conf: SQLConf) extends SqlBaseBaseVisitor[AnyRef] with Logging
    */
   override def visitFunctionCall(ctx: FunctionCallContext): Expression = withOrigin(ctx) {
     // Create the function call.
-    val name = ctx.qualifiedName.getText
+    val name = ctx.functionCallName.getText
     val isDistinct = Option(ctx.setQuantifier()).exists(_.DISTINCT != null)
     val arguments = ctx.argument.asScala.map(expression) match {
       case Seq(UnresolvedStar(None))
@@ -1595,7 +1595,8 @@ class AstBuilder(conf: SQLConf) extends SqlBaseBaseVisitor[AnyRef] with Logging
       case expressions =>
         expressions
     }
-    val function = UnresolvedFunction(visitFunctionName(ctx.qualifiedName), arguments, isDistinct)
+    val function = UnresolvedFunction(
+      visitFunctionName(ctx.functionCallName), arguments, isDistinct)
 
     // Check if the function is evaluated in a windowed context.
     ctx.windowSpec match {
@@ -1623,6 +1624,17 @@ class AstBuilder(conf: SQLConf) extends SqlBaseBaseVisitor[AnyRef] with Logging
     visitFunctionName(ctx, ctx.identifier().asScala.map(_.getText))
   }
 
+  /**
+   * Create a function database (optional) and name pair.
+   */
+  protected def visitFunctionName(ctx: FunctionCallNameContext): FunctionIdentifier = {
+    if (ctx.qualifiedName != null) {
+      visitFunctionName(ctx.qualifiedName)
+    } else {
+      FunctionIdentifier(ctx.getText, None)
+    }
+  }
+
   /**
    * Create a function database (optional) and name pair.
    */
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/ParseDriver.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/ParseDriver.scala
@@ -101,15 +101,15 @@ abstract class AbstractSqlParser(conf: SQLConf) extends ParserInterface with Log
     lexer.removeErrorListeners()
     lexer.addErrorListener(ParseErrorListener)
     lexer.legacy_setops_precedence_enbled = conf.setOpsPrecedenceEnforced
-    lexer.ansi = conf.dialectSparkAnsiEnabled
+    lexer.ansi = conf.ansiEnabled
 
     val tokenStream = new CommonTokenStream(lexer)
     val parser = new SqlBaseParser(tokenStream)
     parser.addParseListener(PostProcessor)
     parser.removeErrorListeners()
     parser.addErrorListener(ParseErrorListener)
     parser.legacy_setops_precedence_enbled = conf.setOpsPrecedenceEnforced
-    parser.ansi = conf.dialectSparkAnsiEnabled
+    parser.ansi = conf.ansiEnabled
 
     try {
       try {
diff --git a/sql/core/src/test/resources/sql-tests/inputs/postgreSQL/text.sql b/sql/core/src/test/resources/sql-tests/inputs/postgreSQL/text.sql
@@ -44,11 +44,7 @@ select concat_ws(',',10,20,null,30);
 select concat_ws('',10,20,null,30);
 select concat_ws(NULL,10,20,null,30) is null;
 select reverse('abcde');
--- [SPARK-28036] Built-in udf left/right has inconsistent behavior
--- [SPARK-28479][SPARK-28989] Parser error when enabling ANSI mode
-set spark.sql.dialect.spark.ansi.enabled=false;
 select i, left('ahoj', i), right('ahoj', i) from range(-5, 6) t(i) order by i;
-set spark.sql.dialect.spark.ansi.enabled=true;
 -- [SPARK-28037] Add built-in String Functions: quote_literal
 -- select quote_literal('');
 -- select quote_literal('abc''');
diff --git a/sql/core/src/test/resources/sql-tests/results/postgreSQL/text.sql.out b/sql/core/src/test/resources/sql-tests/results/postgreSQL/text.sql.out
@@ -1,5 +1,5 @@
 -- Automatically generated by SQLQueryTestSuite
--- Number of queries: 44
+-- Number of queries: 42
 
 
 -- !query 0
@@ -151,18 +151,10 @@ edcba
 
 
 -- !query 18
-set spark.sql.dialect.spark.ansi.enabled=false
--- !query 18 schema
-struct<key:string,value:string>
--- !query 18 output
-spark.sql.dialect.spark.ansi.enabled	false
-
-
--- !query 19
 select i, left('ahoj', i), right('ahoj', i) from range(-5, 6) t(i) order by i
--- !query 19 schema
+-- !query 18 schema
 struct<i:bigint,left('ahoj', t.`i`):string,right('ahoj', t.`i`):string>
--- !query 19 output
+-- !query 18 output
 -5		
 -4		
 -3		
@@ -176,200 +168,192 @@ struct<i:bigint,left('ahoj', t.`i`):string,right('ahoj', t.`i`):string>
 5	ahoj	ahoj
 
 
--- !query 20
-set spark.sql.dialect.spark.ansi.enabled=true
--- !query 20 schema
-struct<key:string,value:string>
--- !query 20 output
-spark.sql.dialect.spark.ansi.enabled	true
-
-
--- !query 21
+-- !query 19
 /*
  * format
  */
 select format_string(NULL)
--- !query 21 schema
+-- !query 19 schema
 struct<format_string(CAST(NULL AS STRING)):string>
--- !query 21 output
+-- !query 19 output
 NULL
 
 
--- !query 22
+-- !query 20
 select format_string('Hello')
--- !query 22 schema
+-- !query 20 schema
 struct<format_string(Hello):string>
--- !query 22 output
+-- !query 20 output
 Hello
 
 
--- !query 23
+-- !query 21
 select format_string('Hello %s', 'World')
--- !query 23 schema
+-- !query 21 schema
 struct<format_string(Hello %s, World):string>
--- !query 23 output
+-- !query 21 output
 Hello World
 
 
--- !query 24
+-- !query 22
 select format_string('Hello %%')
--- !query 24 schema
+-- !query 22 schema
 struct<format_string(Hello %%):string>
--- !query 24 output
+-- !query 22 output
 Hello %
 
 
--- !query 25
+-- !query 23
 select format_string('Hello %%%%')
--- !query 25 schema
+-- !query 23 schema
 struct<format_string(Hello %%%%):string>
--- !query 25 output
+-- !query 23 output
 Hello %%
 
 
--- !query 26
+-- !query 24
 select format_string('Hello %s %s', 'World')
--- !query 26 schema
+-- !query 24 schema
 struct<>
--- !query 26 output
+-- !query 24 output
 java.util.MissingFormatArgumentException
 Format specifier '%s'
 
 
--- !query 27
+-- !query 25
 select format_string('Hello %s')
--- !query 27 schema
+-- !query 25 schema
 struct<>
--- !query 27 output
+-- !query 25 output
 java.util.MissingFormatArgumentException
 Format specifier '%s'
 
 
--- !query 28
+-- !query 26
 select format_string('Hello %x', 20)
--- !query 28 schema
+-- !query 26 schema
 struct<format_string(Hello %x, 20):string>
--- !query 28 output
+-- !query 26 output
 Hello 14
 
 
--- !query 29
+-- !query 27
 select format_string('%1$s %3$s', 1, 2, 3)
--- !query 29 schema
+-- !query 27 schema
 struct<format_string(%1$s %3$s, 1, 2, 3):string>
--- !query 29 output
+-- !query 27 output
 1 3
 
 
--- !query 30
+-- !query 28
 select format_string('%1$s %12$s', 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
--- !query 30 schema
+-- !query 28 schema
 struct<format_string(%1$s %12$s, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12):string>
--- !query 30 output
+-- !query 28 output
 1 12
 
 
--- !query 31
+-- !query 29
 select format_string('%1$s %4$s', 1, 2, 3)
--- !query 31 schema
+-- !query 29 schema
 struct<>
--- !query 31 output
+-- !query 29 output
 java.util.MissingFormatArgumentException
 Format specifier '%4$s'
 
 
--- !query 32
+-- !query 30
 select format_string('%1$s %13$s', 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
--- !query 32 schema
+-- !query 30 schema
 struct<>
--- !query 32 output
+-- !query 30 output
 java.util.MissingFormatArgumentException
 Format specifier '%13$s'
 
 
--- !query 33
+-- !query 31
 select format_string('%0$s', 'Hello')
--- !query 33 schema
+-- !query 31 schema
 struct<format_string(%0$s, Hello):string>
--- !query 33 output
+-- !query 31 output
 Hello
 
 
--- !query 34
+-- !query 32
 select format_string('Hello %s %1$s %s', 'World', 'Hello again')
--- !query 34 schema
+-- !query 32 schema
 struct<format_string(Hello %s %1$s %s, World, Hello again):string>
--- !query 34 output
+-- !query 32 output
 Hello World World Hello again
 
 
--- !query 35
+-- !query 33
 select format_string('Hello %s %s, %2$s %2$s', 'World', 'Hello again')
--- !query 35 schema
+-- !query 33 schema
 struct<format_string(Hello %s %s, %2$s %2$s, World, Hello again):string>
--- !query 35 output
+-- !query 33 output
 Hello World Hello again, Hello again Hello again
 
 
--- !query 36
+-- !query 34
 select format_string('>>%10s<<', 'Hello')
--- !query 36 schema
+-- !query 34 schema
 struct<format_string(>>%10s<<, Hello):string>
--- !query 36 output
+-- !query 34 output
 >>     Hello<<
 
 
--- !query 37
+-- !query 35
 select format_string('>>%10s<<', NULL)
--- !query 37 schema
+-- !query 35 schema
 struct<format_string(>>%10s<<, NULL):string>
--- !query 37 output
+-- !query 35 output
 >>      null<<
 
 
--- !query 38
+-- !query 36
 select format_string('>>%10s<<', '')
--- !query 38 schema
+-- !query 36 schema
 struct<format_string(>>%10s<<, ):string>
--- !query 38 output
+-- !query 36 output
 >>          <<
 
 
--- !query 39
+-- !query 37
 select format_string('>>%-10s<<', '')
--- !query 39 schema
+-- !query 37 schema
 struct<format_string(>>%-10s<<, ):string>
--- !query 39 output
+-- !query 37 output
 >>          <<
 
 
--- !query 40
+-- !query 38
 select format_string('>>%-10s<<', 'Hello')
--- !query 40 schema
+-- !query 38 schema
 struct<format_string(>>%-10s<<, Hello):string>
--- !query 40 output
+-- !query 38 output
 >>Hello     <<
 
 
--- !query 41
+-- !query 39
 select format_string('>>%-10s<<', NULL)
--- !query 41 schema
+-- !query 39 schema
 struct<format_string(>>%-10s<<, NULL):string>
--- !query 41 output
+-- !query 39 output
 >>null      <<
 
 
--- !query 42
+-- !query 40
 select format_string('>>%1$10s<<', 'Hello')
--- !query 42 schema
+-- !query 40 schema
 struct<format_string(>>%1$10s<<, Hello):string>
--- !query 42 output
+-- !query 40 output
 >>     Hello<<
 
 
--- !query 43
+-- !query 41
 DROP TABLE TEXT_TBL
--- !query 43 schema
+-- !query 41 schema
 struct<>
--- !query 43 output
+-- !query 41 output
 
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/SQLQueryTestSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/SQLQueryTestSuite.scala
@@ -343,7 +343,6 @@ class SQLQueryTestSuite extends QueryTest with SharedSparkSession {
         localSparkSession.udf.register("boolne", (b1: Boolean, b2: Boolean) => b1 != b2)
         // vol used by boolean.sql and case.sql.
         localSparkSession.udf.register("vol", (s: String) => s)
-        localSparkSession.conf.set(SQLConf.DIALECT_SPARK_ANSI_ENABLED.key, true)
         localSparkSession.conf.set(SQLConf.DIALECT.key, SQLConf.Dialect.POSTGRESQL.toString)
       case _: AnsiTest =>
         localSparkSession.conf.set(SQLConf.DIALECT_SPARK_ANSI_ENABLED.key, true)
