Initial commit

jovanm-db · Jan 8, 2025 · a0bb4eb · a0bb4eb
1 parent 82ef90c
commit a0bb4eb
Show file tree

Hide file tree

Showing 4 changed files with 33 additions and 1 deletion.
diff --git a/sql/api/src/main/scala/org/apache/spark/sql/internal/types/AbstractStringType.scala b/sql/api/src/main/scala/org/apache/spark/sql/internal/types/AbstractStringType.scala
@@ -17,7 +17,7 @@
 
 package org.apache.spark.sql.internal.types
 
-import org.apache.spark.sql.types.{AbstractDataType, DataType, StringType}
+import org.apache.spark.sql.types.{AbstractDataType, CharType, DataType, StringType, VarcharType}
 
 /**
  * AbstractStringType is an abstract class for StringType with collation support.
@@ -29,6 +29,8 @@ abstract class AbstractStringType(supportsTrimCollation: Boolean = false)
   override private[sql] def simpleString: String = "string"
 
   override private[sql] def acceptsType(other: DataType): Boolean = other match {
+    case CharType(_) | VarcharType(_) =>
+      false
     case st: StringType =>
       canUseTrimCollation(st) && acceptsStringType(st)
     case _ =>

diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/AnsiTypeCoercion.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/AnsiTypeCoercion.scala
@@ -177,6 +177,7 @@ object AnsiTypeCoercion extends TypeCoercionBase {
         } else {
           None
         }
+      case (CharType(_) | VarcharType(_), _: AbstractStringType) => Some(StringType)
       case (_: StringType, _: AbstractStringType) => None
 
       // If a function expects integral type, fractional input is not allowed.

diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/TypeCoercion.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/TypeCoercion.scala
@@ -202,6 +202,7 @@ object TypeCoercion extends TypeCoercionBase {
         } else {
           null
         }
+      case (CharType(_) | VarcharType(_), _: AbstractStringType) => StringType
       // If the function accepts any numeric type and the input is a string, we follow the hive
       // convention and cast that input into a double
       case (_: StringType, NumericType) => NumericType.defaultConcreteType

diff --git a/sql/core/src/test/scala/org/apache/spark/sql/CharVarcharTestSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/CharVarcharTestSuite.scala
@@ -920,6 +920,34 @@ class BasicCharVarcharTestSuite extends QueryTest with SharedSparkSession {
       }
     }
   }
+
+  test("string expressions") {
+    def run(typ: String, expected: Row): Unit = {
+      val df = sql(
+      s"""
+         |SELECT
+         |'X' || '5 '::$typ(5) || 'X',
+         |LENGTH('5 '::$typ(5)),
+         |UPPER('a '::$typ(5)),
+         |LENGTH(UPPER('a '::$typ(5))),
+         |'"' || UPPER('a '::$typ(5)) || '"',
+         |REPLACE('abc', 'ab '::$typ(5), 'x'),
+         |TRANSLATE('abc', 'abc', 'xy '::$typ(5))
+      """.stripMargin)
+      assert(df.schema.map(_.dataType) == Seq(StringType, IntegerType, StringType, IntegerType,
+        StringType, StringType, StringType))
+      assert(df.collect() === Array(expected))
+    }
+
+    val expected = Row("X5 X", 2, "A ", 2, "\"A \"", "abc", "xy ")
+    Seq(("char", expected, Row("X5    X", 5, "A    ", 5, "\"A    \"", "abc", "xy ")),
+      ("varchar", expected, expected)).foreach { case (typ, a, b) =>
+      run(typ, a)
+      withSQLConf((SQLConf.PRESERVE_CHAR_VARCHAR_TYPE_INFO.key, "true")) {
+        run(typ, b)
+      }
+    }
+  }
 }
 
 class FileSourceCharVarcharTestSuite extends CharVarcharTestSuite with SharedSparkSession {