introduce field validator

typesystem/src/main/scala/org/apache/hadoop/metadata/query/QueryProcessor.scala

@@ -48,7 +48,15 @@ object QueryProcessor {
      */
     e1.dataType
 
-    e1
+    /*
+     * ensure fieldReferences match the input expression's dataType
+     */
+    val e2 = e1.transformUp(FieldValidator)
+    val e3 = e2.transformUp(new Resolver())
+
+    e3.dataType
+
+    e3
   }
 
 }

typesystem/src/main/scala/org/apache/hadoop/metadata/query/Resolver.scala

@@ -19,6 +19,7 @@
 package org.apache.hadoop.metadata.query
 
 import Expressions._
+import org.apache.hadoop.metadata.types.IDataType
 
 class Resolver(srcExpr : Option[Expression] = None, aliases : Map[String, Expression] = Map())
 extends PartialFunction[Expression, Expression] {
@@ -75,3 +76,43 @@ extends PartialFunction[Expression, Expression] {
     case x => x
   }
 }
+
+/**
+ * - any FieldReferences that explicitly reference the input, can be converted to implicit references
+ * - any FieldReferences that explicitly reference a
+ */
+object FieldValidator extends PartialFunction[Expression, Expression] {
+
+  def isDefinedAt(x: Expression) = true
+
+  def validateQualifiedField(srcDataType : IDataType[_]) : PartialFunction[Expression, Expression] = {
+    case FieldExpression(fNm, fInfo, Some(child)) if child.children == Nil && child.dataType == srcDataType =>
+      FieldExpression(fNm, fInfo, None)
+    case fe@FieldExpression(fNm, fInfo, Some(child)) if child.children == Nil =>
+      throw new ExpressionException(fe, s"srcType of field doesn't match input type")
+    case hasFieldUnaryExpression(fNm, child) if child.dataType == srcDataType =>
+      hasFieldLeafExpression(fNm)
+    case hF@hasFieldUnaryExpression(fNm, child) =>
+      throw new ExpressionException(hF, s"srcType of field doesn't match input type")
+    case isTraitUnaryExpression(fNm, child) if child.dataType == srcDataType =>
+      isTraitLeafExpression(fNm)
+    case iT@isTraitUnaryExpression(fNm, child) =>
+      throw new ExpressionException(iT, s"srcType of field doesn't match input type")
+  }
+
+  def apply(e : Expression) : Expression = e match {
+    case f@FilterExpression(inputExpr, condExpr) => {
+      val validatedCE = condExpr.transformUp(validateQualifiedField(inputExpr.dataType))
+      if ( validatedCE.fastEquals(condExpr) ) {
+        f
+      } else {
+        new FilterExpression(inputExpr, validatedCE)
+      }
+    }
+    case SelectExpression(child, selectList) if child.resolved => {
+      val v = validateQualifiedField(child.dataType)
+      return new SelectExpression(child, selectList.map{_.transformUp(v)})
+    }
+    case x => x
+  }
+}
\ No newline at end of file

typesystem/src/test/scala/org/apache/hadoop/metadata/query/ExpressionTest.scala

@@ -92,6 +92,21 @@ class ExpressionTest extends BaseTest {
     println(e)
   }
 
+  @Test def testNegFieldReference: Unit = {
+    try {
+      val e = QueryProcessor.validate(_class("DB").where(_class("LoadProcess").hasField("name")))
+      println(e)
+    } catch {
+      case e : ExpressionException
+        if e.getMessage.endsWith("srcType of field doesn't match input type, expression: LoadProcess has name") => ()
+    }
+  }
+
+  @Test def testFieldReferenceRedundant: Unit = {
+    val e = QueryProcessor.validate(_class("DB").where(_class("DB").hasField("name")))
+    println(e)
+  }
+
   @Test def testBackReference: Unit = {
     val e = QueryProcessor.validate(
       _class("DB").as("db").field("Table").where(id("db").field("name").`=`(string("Reporting"))))

typesystem/src/test/scala/org/apache/hadoop/metadata/query/GremlinTest.scala

@@ -51,6 +51,12 @@ class GremlinTest extends FunSuite with BeforeAndAfterAll {
     validateJson(r, "{\n  \"query\":\"DB where (name = \\\"Reporting\\\")\",\n  \"dataType\":{\n    \"superTypes\":[\n      \n    ],\n    \"hierarchicalMetaTypeName\":\"org.apache.hadoop.metadata.types.ClassType\",\n    \"typeName\":\"DB\",\n    \"attributeDefinitions\":[\n      {\n        \"name\":\"name\",\n        \"dataTypeName\":\"string\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      },\n      {\n        \"name\":\"owner\",\n        \"dataTypeName\":\"string\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      },\n      {\n        \"name\":\"createTime\",\n        \"dataTypeName\":\"int\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      }\n    ]\n  },\n  \"rows\":[\n    {\n      \"$typeName$\":\"DB\",\n      \"$id$\":{\n        \"id\":\"7168\",\n        \"$typeName$\":\"DB\",\n        \"version\":0\n      },\n      \"owner\":\"Jane BI\",\n      \"name\":\"Reporting\",\n      \"createTime\":1500\n    }\n  ]\n}")
   }
 
+  test("testFilter2") {
+    var r = QueryProcessor.evaluate(_class("DB").where(id("DB").field("name").`=`(string("Reporting"))), g)
+    validateJson(r, "{\n  \"query\":\"DB where (name = \\\"Reporting\\\")\",\n  \"dataType\":{\n    \"superTypes\":[\n      \n    ],\n    \"hierarchicalMetaTypeName\":\"org.apache.hadoop.metadata.types.ClassType\",\n    \"typeName\":\"DB\",\n    \"attributeDefinitions\":[\n      {\n        \"name\":\"name\",\n        \"dataTypeName\":\"string\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      },\n      {\n        \"name\":\"owner\",\n        \"dataTypeName\":\"string\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      },\n      {\n        \"name\":\"createTime\",\n        \"dataTypeName\":\"int\",\n        \"multiplicity\":{\n          \"lower\":0,\n          \"upper\":1,\n          \"isUnique\":false\n        },\n        \"isComposite\":false,\n        \"isUnique\":false,\n        \"isIndexable\":true,\n        \"reverseAttributeName\":null\n      }\n    ]\n  },\n  \"rows\":[\n    {\n      \"$typeName$\":\"DB\",\n      \"$id$\":{\n        \"id\":\"7168\",\n        \"$typeName$\":\"DB\",\n        \"version\":0\n      },\n      \"owner\":\"Jane BI\",\n      \"name\":\"Reporting\",\n      \"createTime\":1500\n    }\n  ]\n}")
+  }
+
+
   test("testSelect") {
     val r = QueryProcessor.evaluate(_class("DB").where(id("name").`=`(string("Reporting"))).
       select(id("name"), id("owner")), g)

typesystem/src/test/scala/org/apache/hadoop/metadata/query/QueryTestsUtils.scala

@@ -92,7 +92,7 @@ object QueryTestsUtils extends GraphUtils {
     def loadProcessClsDef = new HierarchicalTypeDefinition[ClassType](classOf[ClassType], "LoadProcess", null,
       Array(
         attrDef("name", DataTypes.STRING_TYPE),
-        new AttributeDefinition("inputTables", "Table", Multiplicity.COLLECTION, false, null),
+        new AttributeDefinition("inputTables", DataTypes.arrayTypeName("Table"), Multiplicity.COLLECTION, false, null),
         new AttributeDefinition("outputTable", "Table", Multiplicity.REQUIRED, false, null)
       ))