clean up hivesterix
git-svn-id: https://hyracks.googlecode.com/svn/branches/fullstack_job_rewriter@3008 123451ca-8445-de46-9d55-352943316053
diff --git a/hivesterix/pom.xml b/hivesterix/pom.xml
index cf9c4b5..55f837e 100644
--- a/hivesterix/pom.xml
+++ b/hivesterix/pom.xml
@@ -4,7 +4,7 @@
<groupId>edu.uci.ics.hivesterix</groupId>
<artifactId>hivesterix</artifactId>
<version>0.2.3-SNAPSHOT</version>
- <name>hivesterix</name>
+ <name>hivesterix</name>
<dependencies>
<dependency>
<groupId>javax.servlet</groupId>
@@ -313,6 +313,25 @@
</configuration>
</plugin>
<plugin>
+ <artifactId>maven-jar-plugin</artifactId>
+ <executions>
+ <execution>
+ <id>patch</id>
+ <goals>
+ <goal>jar</goal>
+ </goals>
+ <phase>package</phase>
+ <configuration>
+ <classifier>patch</classifier>
+ <finalName>a-hive</finalName>
+ <includes>
+ <include>**/org/apache/**</include>
+ </includes>
+ </configuration>
+ </execution>
+ </executions>
+ </plugin>
+ <plugin>
<groupId>org.codehaus.mojo</groupId>
<artifactId>appassembler-maven-plugin</artifactId>
<version>1.3</version>
@@ -358,8 +377,8 @@
<version>2.13</version>
<configuration>
<forkMode>pertest</forkMode>
- <argLine>-enableassertions -Xmx2047m -Dfile.encoding=UTF-8
- -Djava.util.logging.config.file=src/test/resources/logging.properties</argLine>
+ <argLine>-enableassertions -Xmx2047m -Dfile.encoding=UTF-8
+ -Djava.util.logging.config.file=src/test/resources/logging.properties</argLine>
<includes>
<include>**/test/optimizer/*TestSuite.java</include>
<include>**/test/optimizer/*Test.java</include>
diff --git a/hivesterix/src/main/assembly/binary-assembly.xml b/hivesterix/src/main/assembly/binary-assembly.xml
index 0500499..906f199 100755
--- a/hivesterix/src/main/assembly/binary-assembly.xml
+++ b/hivesterix/src/main/assembly/binary-assembly.xml
@@ -1,19 +1,26 @@
<assembly>
- <id>binary-assembly</id>
- <formats>
- <format>zip</format>
- <format>dir</format>
- </formats>
- <includeBaseDirectory>false</includeBaseDirectory>
- <fileSets>
- <fileSet>
- <directory>target/appassembler/bin</directory>
- <outputDirectory>bin</outputDirectory>
- <fileMode>0755</fileMode>
- </fileSet>
- <fileSet>
- <directory>target/appassembler/lib</directory>
- <outputDirectory>lib</outputDirectory>
- </fileSet>
- </fileSets>
+ <id>binary-assembly</id>
+ <formats>
+ <format>zip</format>
+ <format>dir</format>
+ </formats>
+ <includeBaseDirectory>false</includeBaseDirectory>
+ <fileSets>
+ <fileSet>
+ <directory>target/appassembler/bin</directory>
+ <outputDirectory>bin</outputDirectory>
+ <fileMode>0755</fileMode>
+ </fileSet>
+ <fileSet>
+ <directory>target/appassembler/lib</directory>
+ <outputDirectory>lib</outputDirectory>
+ </fileSet>
+ <fileSet>
+ <directory>target</directory>
+ <outputDirectory>lib</outputDirectory>
+ <includes>
+ <include>**/*.jar</include>
+ </includes>
+ </fileSet>
+ </fileSets>
</assembly>
diff --git a/hivesterix/src/main/java/edu/uci/ics/hivesterix/logical/plan/HiveAlgebricksTranslator.java b/hivesterix/src/main/java/edu/uci/ics/hivesterix/logical/plan/HiveAlgebricksTranslator.java
index 1fb973e..2765e44 100644
--- a/hivesterix/src/main/java/edu/uci/ics/hivesterix/logical/plan/HiveAlgebricksTranslator.java
+++ b/hivesterix/src/main/java/edu/uci/ics/hivesterix/logical/plan/HiveAlgebricksTranslator.java
@@ -80,729 +80,770 @@
@SuppressWarnings("rawtypes")
public class HiveAlgebricksTranslator implements Translator {
- private int currentVariable = 0;
+ private int currentVariable = 0;
- private List<Mutable<ILogicalOperator>> logicalOp = new ArrayList<Mutable<ILogicalOperator>>();
+ private List<Mutable<ILogicalOperator>> logicalOp = new ArrayList<Mutable<ILogicalOperator>>();
- private boolean continueTraverse = true;
+ private boolean continueTraverse = true;
- private IMetadataProvider<PartitionDesc, Object> metaData;
+ private IMetadataProvider<PartitionDesc, Object> metaData;
- /**
- * map variable name to the logical variable
- */
- private HashMap<String, LogicalVariable> nameToLogicalVariableMap = new HashMap<String, LogicalVariable>();
+ /**
+ * map variable name to the logical variable
+ */
+ private HashMap<String, LogicalVariable> nameToLogicalVariableMap = new HashMap<String, LogicalVariable>();
- /**
- * map field name to LogicalVariable
- */
- private HashMap<String, LogicalVariable> fieldToLogicalVariableMap = new HashMap<String, LogicalVariable>();
+ /**
+ * map field name to LogicalVariable
+ */
+ private HashMap<String, LogicalVariable> fieldToLogicalVariableMap = new HashMap<String, LogicalVariable>();
- /**
- * map logical variable to name
- */
- private HashMap<LogicalVariable, String> logicalVariableToFieldMap = new HashMap<LogicalVariable, String>();
+ /**
+ * map logical variable to name
+ */
+ private HashMap<LogicalVariable, String> logicalVariableToFieldMap = new HashMap<LogicalVariable, String>();
- /**
- * asterix root operators
- */
- private List<Mutable<ILogicalOperator>> rootOperators = new ArrayList<Mutable<ILogicalOperator>>();
+ /**
+ * asterix root operators
+ */
+ private List<Mutable<ILogicalOperator>> rootOperators = new ArrayList<Mutable<ILogicalOperator>>();
- /**
- * a list of visitors
- */
- private List<Visitor> visitors = new ArrayList<Visitor>();
+ /**
+ * a list of visitors
+ */
+ private List<Visitor> visitors = new ArrayList<Visitor>();
- /**
- * output writer to print things out
- */
- private static PrintWriter outputWriter = new PrintWriter(new OutputStreamWriter(System.out));
+ /**
+ * output writer to print things out
+ */
+ private static PrintWriter outputWriter = new PrintWriter(
+ new OutputStreamWriter(System.out));
- /**
- * map a logical variable to type info
- */
- private HashMap<LogicalVariable, TypeInfo> variableToType = new HashMap<LogicalVariable, TypeInfo>();
+ /**
+ * map a logical variable to type info
+ */
+ private HashMap<LogicalVariable, TypeInfo> variableToType = new HashMap<LogicalVariable, TypeInfo>();
- @Override
- public LogicalVariable getVariable(String fieldName, TypeInfo type) {
- LogicalVariable var = fieldToLogicalVariableMap.get(fieldName);
- if (var == null) {
- currentVariable++;
- var = new LogicalVariable(currentVariable);
- fieldToLogicalVariableMap.put(fieldName, var);
- nameToLogicalVariableMap.put(var.toString(), var);
- variableToType.put(var, type);
- logicalVariableToFieldMap.put(var, fieldName);
- }
- return var;
- }
+ @Override
+ public LogicalVariable getVariable(String fieldName, TypeInfo type) {
+ LogicalVariable var = fieldToLogicalVariableMap.get(fieldName);
+ if (var == null) {
+ currentVariable++;
+ var = new LogicalVariable(currentVariable);
+ fieldToLogicalVariableMap.put(fieldName, var);
+ nameToLogicalVariableMap.put(var.toString(), var);
+ variableToType.put(var, type);
+ logicalVariableToFieldMap.put(var, fieldName);
+ }
+ return var;
+ }
- @Override
- public LogicalVariable getNewVariable(String fieldName, TypeInfo type) {
- currentVariable++;
- LogicalVariable var = new LogicalVariable(currentVariable);
- fieldToLogicalVariableMap.put(fieldName, var);
- nameToLogicalVariableMap.put(var.toString(), var);
- variableToType.put(var, type);
- logicalVariableToFieldMap.put(var, fieldName);
- return var;
- }
+ @Override
+ public LogicalVariable getNewVariable(String fieldName, TypeInfo type) {
+ currentVariable++;
+ LogicalVariable var = new LogicalVariable(currentVariable);
+ fieldToLogicalVariableMap.put(fieldName, var);
+ nameToLogicalVariableMap.put(var.toString(), var);
+ variableToType.put(var, type);
+ logicalVariableToFieldMap.put(var, fieldName);
+ return var;
+ }
- @Override
- public void replaceVariable(LogicalVariable oldVar, LogicalVariable newVar) {
- String name = this.logicalVariableToFieldMap.get(oldVar);
- if (name != null) {
- fieldToLogicalVariableMap.put(name, newVar);
- nameToLogicalVariableMap.put(newVar.toString(), newVar);
- nameToLogicalVariableMap.put(oldVar.toString(), newVar);
- logicalVariableToFieldMap.put(newVar, name);
- }
- }
+ @Override
+ public void replaceVariable(LogicalVariable oldVar, LogicalVariable newVar) {
+ String name = this.logicalVariableToFieldMap.get(oldVar);
+ if (name != null) {
+ fieldToLogicalVariableMap.put(name, newVar);
+ nameToLogicalVariableMap.put(newVar.toString(), newVar);
+ nameToLogicalVariableMap.put(oldVar.toString(), newVar);
+ logicalVariableToFieldMap.put(newVar, name);
+ }
+ }
- @Override
- public IMetadataProvider<PartitionDesc, Object> getMetadataProvider() {
- return metaData;
- }
+ @Override
+ public IMetadataProvider<PartitionDesc, Object> getMetadataProvider() {
+ return metaData;
+ }
- /**
- * only get an variable, without rewriting it
- *
- * @param fieldName
- * @return
- */
- private LogicalVariable getVariableOnly(String fieldName) {
- return fieldToLogicalVariableMap.get(fieldName);
- }
+ /**
+ * only get an variable, without rewriting it
+ *
+ * @param fieldName
+ * @return
+ */
+ private LogicalVariable getVariableOnly(String fieldName) {
+ return fieldToLogicalVariableMap.get(fieldName);
+ }
- private void updateVariable(String fieldName, LogicalVariable variable) {
- LogicalVariable var = fieldToLogicalVariableMap.get(fieldName);
- if (var == null) {
- fieldToLogicalVariableMap.put(fieldName, variable);
- nameToLogicalVariableMap.put(fieldName, variable);
- } else if (!var.equals(variable)) {
- // System.out.println("!!!replace variables!!!");
- fieldToLogicalVariableMap.put(fieldName, variable);
- nameToLogicalVariableMap.put(fieldName, variable);
- }
- }
+ private void updateVariable(String fieldName, LogicalVariable variable) {
+ LogicalVariable var = fieldToLogicalVariableMap.get(fieldName);
+ if (var == null) {
+ fieldToLogicalVariableMap.put(fieldName, variable);
+ nameToLogicalVariableMap.put(fieldName, variable);
+ } else if (!var.equals(variable)) {
+ fieldToLogicalVariableMap.put(fieldName, variable);
+ nameToLogicalVariableMap.put(fieldName, variable);
+ }
+ }
- /**
- * get a list of logical variables from the schema
- *
- * @param schema
- * @return
- */
- @Override
- public List<LogicalVariable> getVariablesFromSchema(Schema schema) {
- List<LogicalVariable> variables = new ArrayList<LogicalVariable>();
- List<String> names = schema.getNames();
+ /**
+ * get a list of logical variables from the schema
+ *
+ * @param schema
+ * @return
+ */
+ @Override
+ public List<LogicalVariable> getVariablesFromSchema(Schema schema) {
+ List<LogicalVariable> variables = new ArrayList<LogicalVariable>();
+ List<String> names = schema.getNames();
- for (String name : names)
- variables.add(nameToLogicalVariableMap.get(name));
- return variables;
- }
+ for (String name : names)
+ variables.add(nameToLogicalVariableMap.get(name));
+ return variables;
+ }
- /**
- * get variable to typeinfo map
- *
- * @return
- */
- public HashMap<LogicalVariable, TypeInfo> getVariableContext() {
- return this.variableToType;
- }
+ /**
+ * get variable to typeinfo map
+ *
+ * @return
+ */
+ public HashMap<LogicalVariable, TypeInfo> getVariableContext() {
+ return this.variableToType;
+ }
- /**
- * get the number of variables
- * s
- *
- * @return
- */
- public int getVariableCounter() {
- return currentVariable + 1;
- }
+ /**
+ * get the number of variables s
+ *
+ * @return
+ */
+ public int getVariableCounter() {
+ return currentVariable + 1;
+ }
- /**
- * translate from hive operator tree to asterix operator tree
- *
- * @param hive
- * roots
- * @return Algebricks roots
- */
- public void translate(List<Operator> hiveRoot, ILogicalOperator parentOperator,
- HashMap<String, PartitionDesc> aliasToPathMap) throws AlgebricksException {
- /**
- * register visitors
- */
- visitors.add(new FilterVisitor());
- visitors.add(new GroupByVisitor());
- visitors.add(new JoinVisitor());
- visitors.add(new LateralViewJoinVisitor());
- visitors.add(new UnionVisitor());
- visitors.add(new LimitVisitor());
- visitors.add(new MapJoinVisitor());
- visitors.add(new ProjectVisitor());
- visitors.add(new SortVisitor());
- visitors.add(new ExtractVisitor());
- visitors.add(new TableScanWriteVisitor(aliasToPathMap));
+ /**
+ * translate from hive operator tree to asterix operator tree
+ *
+ * @param hive
+ * roots
+ * @return Algebricks roots
+ */
+ public void translate(List<Operator> hiveRoot,
+ ILogicalOperator parentOperator,
+ HashMap<String, PartitionDesc> aliasToPathMap)
+ throws AlgebricksException {
+ /**
+ * register visitors
+ */
+ visitors.add(new FilterVisitor());
+ visitors.add(new GroupByVisitor());
+ visitors.add(new JoinVisitor());
+ visitors.add(new LateralViewJoinVisitor());
+ visitors.add(new UnionVisitor());
+ visitors.add(new LimitVisitor());
+ visitors.add(new MapJoinVisitor());
+ visitors.add(new ProjectVisitor());
+ visitors.add(new SortVisitor());
+ visitors.add(new ExtractVisitor());
+ visitors.add(new TableScanWriteVisitor(aliasToPathMap));
- List<Mutable<ILogicalOperator>> refList = translate(hiveRoot, new MutableObject<ILogicalOperator>(
- parentOperator));
- insertReplicateOperator(refList);
- if (refList != null)
- rootOperators.addAll(refList);
- }
+ List<Mutable<ILogicalOperator>> refList = translate(hiveRoot,
+ new MutableObject<ILogicalOperator>(parentOperator));
+ insertReplicateOperator(refList);
+ if (refList != null)
+ rootOperators.addAll(refList);
+ }
- /**
- * translate operator DAG
- *
- * @param hiveRoot
- * @param AlgebricksParentOperator
- * @return
- */
- private List<Mutable<ILogicalOperator>> translate(List<Operator> hiveRoot,
- Mutable<ILogicalOperator> AlgebricksParentOperator) throws AlgebricksException {
+ /**
+ * translate operator DAG
+ *
+ * @param hiveRoot
+ * @param AlgebricksParentOperator
+ * @return
+ */
+ private List<Mutable<ILogicalOperator>> translate(List<Operator> hiveRoot,
+ Mutable<ILogicalOperator> AlgebricksParentOperator)
+ throws AlgebricksException {
- for (Operator hiveOperator : hiveRoot) {
- continueTraverse = true;
- Mutable<ILogicalOperator> currentOperatorRef = null;
- if (hiveOperator.getType() == OperatorType.FILTER) {
- FilterOperator fop = (FilterOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.REDUCESINK) {
- ReduceSinkOperator fop = (ReduceSinkOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.JOIN) {
- JoinOperator fop = (JoinOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null) {
- continueTraverse = true;
- break;
- } else
- continueTraverse = false;
- }
- if (currentOperatorRef == null)
- return null;
- } else if (hiveOperator.getType() == OperatorType.LATERALVIEWJOIN) {
- LateralViewJoinOperator fop = (LateralViewJoinOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- if (currentOperatorRef == null)
- return null;
- } else if (hiveOperator.getType() == OperatorType.MAPJOIN) {
- MapJoinOperator fop = (MapJoinOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null) {
- continueTraverse = true;
- break;
- } else
- continueTraverse = false;
- }
- if (currentOperatorRef == null)
- return null;
- } else if (hiveOperator.getType() == OperatorType.SELECT) {
- SelectOperator fop = (SelectOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.EXTRACT) {
- ExtractOperator fop = (ExtractOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.GROUPBY) {
- GroupByOperator fop = (GroupByOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.TABLESCAN) {
- TableScanOperator fop = (TableScanOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.FILESINK) {
- FileSinkOperator fop = (FileSinkOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(fop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.LIMIT) {
- LimitOperator lop = (LimitOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(lop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.UDTF) {
- UDTFOperator lop = (UDTFOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(lop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null)
- break;
- }
- } else if (hiveOperator.getType() == OperatorType.UNION) {
- UnionOperator lop = (UnionOperator) hiveOperator;
- for (Visitor visitor : visitors) {
- currentOperatorRef = visitor.visit(lop, AlgebricksParentOperator, this);
- if (currentOperatorRef != null) {
- continueTraverse = true;
- break;
- } else
- continueTraverse = false;
- }
- } else
- ;
- if (hiveOperator.getChildOperators() != null && hiveOperator.getChildOperators().size() > 0
- && continueTraverse) {
- @SuppressWarnings("unchecked")
- List<Operator> children = hiveOperator.getChildOperators();
- if (currentOperatorRef == null)
- currentOperatorRef = AlgebricksParentOperator;
- translate(children, currentOperatorRef);
- }
- if (hiveOperator.getChildOperators() == null || hiveOperator.getChildOperators().size() == 0)
- logicalOp.add(currentOperatorRef);
- }
- return logicalOp;
- }
+ for (Operator hiveOperator : hiveRoot) {
+ continueTraverse = true;
+ Mutable<ILogicalOperator> currentOperatorRef = null;
+ if (hiveOperator.getType() == OperatorType.FILTER) {
+ FilterOperator fop = (FilterOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.REDUCESINK) {
+ ReduceSinkOperator fop = (ReduceSinkOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.JOIN) {
+ JoinOperator fop = (JoinOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null) {
+ continueTraverse = true;
+ break;
+ } else
+ continueTraverse = false;
+ }
+ if (currentOperatorRef == null)
+ return null;
+ } else if (hiveOperator.getType() == OperatorType.LATERALVIEWJOIN) {
+ LateralViewJoinOperator fop = (LateralViewJoinOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ if (currentOperatorRef == null)
+ return null;
+ } else if (hiveOperator.getType() == OperatorType.MAPJOIN) {
+ MapJoinOperator fop = (MapJoinOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null) {
+ continueTraverse = true;
+ break;
+ } else
+ continueTraverse = false;
+ }
+ if (currentOperatorRef == null)
+ return null;
+ } else if (hiveOperator.getType() == OperatorType.SELECT) {
+ SelectOperator fop = (SelectOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.EXTRACT) {
+ ExtractOperator fop = (ExtractOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.GROUPBY) {
+ GroupByOperator fop = (GroupByOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.TABLESCAN) {
+ TableScanOperator fop = (TableScanOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.FILESINK) {
+ FileSinkOperator fop = (FileSinkOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(fop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.LIMIT) {
+ LimitOperator lop = (LimitOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(lop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.UDTF) {
+ UDTFOperator lop = (UDTFOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(lop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null)
+ break;
+ }
+ } else if (hiveOperator.getType() == OperatorType.UNION) {
+ UnionOperator lop = (UnionOperator) hiveOperator;
+ for (Visitor visitor : visitors) {
+ currentOperatorRef = visitor.visit(lop,
+ AlgebricksParentOperator, this);
+ if (currentOperatorRef != null) {
+ continueTraverse = true;
+ break;
+ } else
+ continueTraverse = false;
+ }
+ } else
+ ;
+ if (hiveOperator.getChildOperators() != null
+ && hiveOperator.getChildOperators().size() > 0
+ && continueTraverse) {
+ @SuppressWarnings("unchecked")
+ List<Operator> children = hiveOperator.getChildOperators();
+ if (currentOperatorRef == null)
+ currentOperatorRef = AlgebricksParentOperator;
+ translate(children, currentOperatorRef);
+ }
+ if (hiveOperator.getChildOperators() == null
+ || hiveOperator.getChildOperators().size() == 0)
+ logicalOp.add(currentOperatorRef);
+ }
+ return logicalOp;
+ }
- /**
- * used in select, group by to get no-column-expression columns
- *
- * @param cols
- * @return
- */
- public ILogicalOperator getAssignOperator(Mutable<ILogicalOperator> parent, List<ExprNodeDesc> cols,
- ArrayList<LogicalVariable> variables) {
+ /**
+ * used in select, group by to get no-column-expression columns
+ *
+ * @param cols
+ * @return
+ */
+ public ILogicalOperator getAssignOperator(Mutable<ILogicalOperator> parent,
+ List<ExprNodeDesc> cols, ArrayList<LogicalVariable> variables) {
- ArrayList<Mutable<ILogicalExpression>> expressions = new ArrayList<Mutable<ILogicalExpression>>();
+ ArrayList<Mutable<ILogicalExpression>> expressions = new ArrayList<Mutable<ILogicalExpression>>();
- /**
- * variables to be appended in the assign operator
- */
- ArrayList<LogicalVariable> appendedVariables = new ArrayList<LogicalVariable>();
+ /**
+ * variables to be appended in the assign operator
+ */
+ ArrayList<LogicalVariable> appendedVariables = new ArrayList<LogicalVariable>();
- // one variable can only be assigned once
- for (ExprNodeDesc hiveExpr : cols) {
- rewriteExpression(hiveExpr);
+ // one variable can only be assigned once
+ for (ExprNodeDesc hiveExpr : cols) {
+ rewriteExpression(hiveExpr);
- if (hiveExpr instanceof ExprNodeColumnDesc) {
- ExprNodeColumnDesc desc2 = (ExprNodeColumnDesc) hiveExpr;
- String fieldName = desc2.getTabAlias() + "." + desc2.getColumn();
+ if (hiveExpr instanceof ExprNodeColumnDesc) {
+ ExprNodeColumnDesc desc2 = (ExprNodeColumnDesc) hiveExpr;
+ String fieldName = desc2.getTabAlias() + "."
+ + desc2.getColumn();
- // System.out.println("project expr: " + fieldName);
+ // System.out.println("project expr: " + fieldName);
- if (fieldName.indexOf("$$") < 0) {
- LogicalVariable var = getVariable(fieldName, hiveExpr.getTypeInfo());
- desc2.setColumn(var.toString());
- desc2.setTabAlias("");
- variables.add(var);
- } else {
- LogicalVariable var = nameToLogicalVariableMap.get(desc2.getColumn());
- String name = this.logicalVariableToFieldMap.get(var);
- var = this.getVariableOnly(name);
- variables.add(var);
- }
- } else {
- Mutable<ILogicalExpression> asterixExpr = translateScalarFucntion(hiveExpr);
- expressions.add(asterixExpr);
- LogicalVariable var = getVariable(hiveExpr.getExprString() + asterixExpr.hashCode(),
- hiveExpr.getTypeInfo());
- variables.add(var);
- appendedVariables.add(var);
- }
- }
+ if (fieldName.indexOf("$$") < 0) {
+ LogicalVariable var = getVariable(fieldName,
+ hiveExpr.getTypeInfo());
+ desc2.setColumn(var.toString());
+ desc2.setTabAlias("");
+ variables.add(var);
+ } else {
+ LogicalVariable var = nameToLogicalVariableMap.get(desc2
+ .getColumn());
+ String name = this.logicalVariableToFieldMap.get(var);
+ var = this.getVariableOnly(name);
+ variables.add(var);
+ }
+ } else {
+ Mutable<ILogicalExpression> asterixExpr = translateScalarFucntion(hiveExpr);
+ expressions.add(asterixExpr);
+ LogicalVariable var = getVariable(hiveExpr.getExprString()
+ + asterixExpr.hashCode(), hiveExpr.getTypeInfo());
+ variables.add(var);
+ appendedVariables.add(var);
+ }
+ }
- /**
- * create an assign operator to deal with appending
- */
- ILogicalOperator assignOp = null;
- if (appendedVariables.size() > 0) {
- assignOp = new AssignOperator(appendedVariables, expressions);
- assignOp.getInputs().add(parent);
- }
- return assignOp;
- }
+ /**
+ * create an assign operator to deal with appending
+ */
+ ILogicalOperator assignOp = null;
+ if (appendedVariables.size() > 0) {
+ assignOp = new AssignOperator(appendedVariables, expressions);
+ assignOp.getInputs().add(parent);
+ }
+ return assignOp;
+ }
- private ILogicalPlan plan;
+ private ILogicalPlan plan;
- public ILogicalPlan genLogicalPlan() {
- plan = new ALogicalPlanImpl(rootOperators);
- return plan;
- }
+ public ILogicalPlan genLogicalPlan() {
+ plan = new ALogicalPlanImpl(rootOperators);
+ return plan;
+ }
- public void printOperators() throws AlgebricksException {
- LogicalOperatorPrettyPrintVisitor pvisitor = new LogicalOperatorPrettyPrintVisitor();
- StringBuilder buffer = new StringBuilder();
- PlanPrettyPrinter.printPlan(plan, buffer, pvisitor, 0);
- outputWriter.println(buffer);
- outputWriter.println("rewritten variables: ");
- outputWriter.flush();
- printVariables();
+ public void printOperators() throws AlgebricksException {
+ LogicalOperatorPrettyPrintVisitor pvisitor = new LogicalOperatorPrettyPrintVisitor();
+ StringBuilder buffer = new StringBuilder();
+ PlanPrettyPrinter.printPlan(plan, buffer, pvisitor, 0);
+ outputWriter.println(buffer);
+ outputWriter.println("rewritten variables: ");
+ outputWriter.flush();
+ printVariables();
- }
+ }
- public static void setOutputPrinter(PrintWriter writer) {
- outputWriter = writer;
- }
+ public static void setOutputPrinter(PrintWriter writer) {
+ outputWriter = writer;
+ }
- private void printVariables() {
- Set<Entry<String, LogicalVariable>> entries = fieldToLogicalVariableMap.entrySet();
+ private void printVariables() {
+ Set<Entry<String, LogicalVariable>> entries = fieldToLogicalVariableMap
+ .entrySet();
- for (Entry<String, LogicalVariable> entry : entries) {
- outputWriter.println(entry.getKey() + " -> " + entry.getValue());
- }
- outputWriter.flush();
- }
+ for (Entry<String, LogicalVariable> entry : entries) {
+ outputWriter.println(entry.getKey() + " -> " + entry.getValue());
+ }
+ outputWriter.flush();
+ }
- /**
- * generate the object inspector for the output of an operator
- *
- * @param operator
- * The Hive operator
- * @return an ObjectInspector object
- */
- public Schema generateInputSchema(Operator operator) {
- List<String> variableNames = new ArrayList<String>();
- List<TypeInfo> typeList = new ArrayList<TypeInfo>();
- List<ColumnInfo> columns = operator.getSchema().getSignature();
+ /**
+ * generate the object inspector for the output of an operator
+ *
+ * @param operator
+ * The Hive operator
+ * @return an ObjectInspector object
+ */
+ public Schema generateInputSchema(Operator operator) {
+ List<String> variableNames = new ArrayList<String>();
+ List<TypeInfo> typeList = new ArrayList<TypeInfo>();
+ List<ColumnInfo> columns = operator.getSchema().getSignature();
- for (ColumnInfo col : columns) {
- // typeList.add();
- TypeInfo type = col.getType();
- typeList.add(type);
+ for (ColumnInfo col : columns) {
+ // typeList.add();
+ TypeInfo type = col.getType();
+ typeList.add(type);
- String fieldName = col.getInternalName();
- variableNames.add(fieldName);
- }
+ String fieldName = col.getInternalName();
+ variableNames.add(fieldName);
+ }
- return new Schema(variableNames, typeList);
- }
+ return new Schema(variableNames, typeList);
+ }
- /**
- * rewrite the names of output columns for feature expression evaluators to
- * use
- *
- * @param operator
- */
- public void rewriteOperatorOutputSchema(Operator operator) {
- List<ColumnInfo> columns = operator.getSchema().getSignature();
+ /**
+ * rewrite the names of output columns for feature expression evaluators to
+ * use
+ *
+ * @param operator
+ */
+ public void rewriteOperatorOutputSchema(Operator operator) {
+ List<ColumnInfo> columns = operator.getSchema().getSignature();
- for (ColumnInfo column : columns) {
- String columnName = column.getTabAlias() + "." + column.getInternalName();
- if (columnName.indexOf("$$") < 0) {
- LogicalVariable var = getVariable(columnName, column.getType());
- column.setInternalName(var.toString());
- }
- }
- }
+ for (ColumnInfo column : columns) {
+ String columnName = column.getTabAlias() + "."
+ + column.getInternalName();
+ if (columnName.indexOf("$$") < 0) {
+ LogicalVariable var = getVariable(columnName, column.getType());
+ column.setInternalName(var.toString());
+ }
+ }
+ }
- @Override
- public void rewriteOperatorOutputSchema(List<LogicalVariable> variables, Operator operator) {
+ @Override
+ public void rewriteOperatorOutputSchema(List<LogicalVariable> variables,
+ Operator operator) {
- printOperatorSchema(operator);
- List<ColumnInfo> columns = operator.getSchema().getSignature();
- if (variables.size() != columns.size()) {
- throw new IllegalStateException("output cardinality error " + operator.getName() + " variable size: "
- + variables.size() + " expected " + columns.size());
- }
+ //printOperatorSchema(operator);
+ List<ColumnInfo> columns = operator.getSchema().getSignature();
+ if (variables.size() != columns.size()) {
+ throw new IllegalStateException("output cardinality error "
+ + operator.getName() + " variable size: "
+ + variables.size() + " expected " + columns.size());
+ }
- for (int i = 0; i < variables.size(); i++) {
- LogicalVariable var = variables.get(i);
- ColumnInfo column = columns.get(i);
- String fieldName = column.getTabAlias() + "." + column.getInternalName();
- if (fieldName.indexOf("$$") < 0) {
- updateVariable(fieldName, var);
- column.setInternalName(var.toString());
- }
- }
- printOperatorSchema(operator);
- }
+ for (int i = 0; i < variables.size(); i++) {
+ LogicalVariable var = variables.get(i);
+ ColumnInfo column = columns.get(i);
+ String fieldName = column.getTabAlias() + "."
+ + column.getInternalName();
+ if (fieldName.indexOf("$$") < 0) {
+ updateVariable(fieldName, var);
+ column.setInternalName(var.toString());
+ }
+ }
+ //printOperatorSchema(operator);
+ }
- /**
- * rewrite an expression and substitute variables
- *
- * @param expr
- * hive expression
- */
- public void rewriteExpression(ExprNodeDesc expr) {
- if (expr instanceof ExprNodeColumnDesc) {
- ExprNodeColumnDesc desc = (ExprNodeColumnDesc) expr;
- String fieldName = desc.getTabAlias() + "." + desc.getColumn();
- if (fieldName.indexOf("$$") < 0) {
- LogicalVariable var = getVariableOnly(fieldName);
- if (var == null) {
- fieldName = "." + desc.getColumn();
- var = getVariableOnly(fieldName);
- if (var == null) {
- fieldName = "null." + desc.getColumn();
- var = getVariableOnly(fieldName);
- if (var == null) {
- throw new IllegalStateException(fieldName + " is wrong!!! ");
- }
- }
- }
- String name = this.logicalVariableToFieldMap.get(var);
- var = getVariableOnly(name);
- desc.setColumn(var.toString());
- }
- } else {
- if (expr.getChildren() != null && expr.getChildren().size() > 0) {
- List<ExprNodeDesc> children = expr.getChildren();
- for (ExprNodeDesc desc : children)
- rewriteExpression(desc);
- }
- }
- }
+ /**
+ * rewrite an expression and substitute variables
+ *
+ * @param expr
+ * hive expression
+ */
+ public void rewriteExpression(ExprNodeDesc expr) {
+ if (expr instanceof ExprNodeColumnDesc) {
+ ExprNodeColumnDesc desc = (ExprNodeColumnDesc) expr;
+ String fieldName = desc.getTabAlias() + "." + desc.getColumn();
+ if (fieldName.indexOf("$$") < 0) {
+ LogicalVariable var = getVariableOnly(fieldName);
+ if (var == null) {
+ fieldName = "." + desc.getColumn();
+ var = getVariableOnly(fieldName);
+ if (var == null) {
+ fieldName = "null." + desc.getColumn();
+ var = getVariableOnly(fieldName);
+ if (var == null) {
+ throw new IllegalStateException(fieldName
+ + " is wrong!!! ");
+ }
+ }
+ }
+ String name = this.logicalVariableToFieldMap.get(var);
+ var = getVariableOnly(name);
+ desc.setColumn(var.toString());
+ }
+ } else {
+ if (expr.getChildren() != null && expr.getChildren().size() > 0) {
+ List<ExprNodeDesc> children = expr.getChildren();
+ for (ExprNodeDesc desc : children)
+ rewriteExpression(desc);
+ }
+ }
+ }
- /**
- * rewrite an expression and substitute variables
- *
- * @param expr
- * hive expression
- */
- public void rewriteExpressionPartial(ExprNodeDesc expr) {
- if (expr instanceof ExprNodeColumnDesc) {
- ExprNodeColumnDesc desc = (ExprNodeColumnDesc) expr;
- String fieldName = desc.getTabAlias() + "." + desc.getColumn();
- if (fieldName.indexOf("$$") < 0) {
- LogicalVariable var = getVariableOnly(fieldName);
- desc.setColumn(var.toString());
- }
- } else {
- if (expr.getChildren() != null && expr.getChildren().size() > 0) {
- List<ExprNodeDesc> children = expr.getChildren();
- for (ExprNodeDesc desc : children)
- rewriteExpressionPartial(desc);
- }
- }
- }
+ /**
+ * rewrite an expression and substitute variables
+ *
+ * @param expr
+ * hive expression
+ */
+ public void rewriteExpressionPartial(ExprNodeDesc expr) {
+ if (expr instanceof ExprNodeColumnDesc) {
+ ExprNodeColumnDesc desc = (ExprNodeColumnDesc) expr;
+ String fieldName = desc.getTabAlias() + "." + desc.getColumn();
+ if (fieldName.indexOf("$$") < 0) {
+ LogicalVariable var = getVariableOnly(fieldName);
+ desc.setColumn(var.toString());
+ }
+ } else {
+ if (expr.getChildren() != null && expr.getChildren().size() > 0) {
+ List<ExprNodeDesc> children = expr.getChildren();
+ for (ExprNodeDesc desc : children)
+ rewriteExpressionPartial(desc);
+ }
+ }
+ }
- private void printOperatorSchema(Operator operator) {
- System.out.println(operator.getName());
- List<ColumnInfo> columns = operator.getSchema().getSignature();
- for (ColumnInfo column : columns) {
- System.out.print(column.getTabAlias() + "." + column.getInternalName() + " ");
- }
- System.out.println();
- }
+ // private void printOperatorSchema(Operator operator) {
+ // // System.out.println(operator.getName());
+ // // List<ColumnInfo> columns = operator.getSchema().getSignature();
+ // // for (ColumnInfo column : columns) {
+ // // System.out.print(column.getTabAlias() + "." +
+ // // column.getInternalName() + " ");
+ // // }
+ // // System.out.println();
+ // }
- /**
- * translate scalar function expression
- *
- * @param hiveExpr
- * @return
- */
- public Mutable<ILogicalExpression> translateScalarFucntion(ExprNodeDesc hiveExpr) {
- ILogicalExpression AlgebricksExpr;
+ /**
+ * translate scalar function expression
+ *
+ * @param hiveExpr
+ * @return
+ */
+ public Mutable<ILogicalExpression> translateScalarFucntion(
+ ExprNodeDesc hiveExpr) {
+ ILogicalExpression AlgebricksExpr;
- if (hiveExpr instanceof ExprNodeGenericFuncDesc) {
- List<Mutable<ILogicalExpression>> arguments = new ArrayList<Mutable<ILogicalExpression>>();
- List<ExprNodeDesc> children = hiveExpr.getChildren();
+ if (hiveExpr instanceof ExprNodeGenericFuncDesc) {
+ List<Mutable<ILogicalExpression>> arguments = new ArrayList<Mutable<ILogicalExpression>>();
+ List<ExprNodeDesc> children = hiveExpr.getChildren();
- for (ExprNodeDesc child : children)
- arguments.add(translateScalarFucntion(child));
+ for (ExprNodeDesc child : children)
+ arguments.add(translateScalarFucntion(child));
- ExprNodeGenericFuncDesc funcExpr = (ExprNodeGenericFuncDesc) hiveExpr;
- GenericUDF genericUdf = funcExpr.getGenericUDF();
- UDF udf = null;
- if (genericUdf instanceof GenericUDFBridge) {
- GenericUDFBridge bridge = (GenericUDFBridge) genericUdf;
- try {
- udf = bridge.getUdfClass().newInstance();
- } catch (Exception e) {
- e.printStackTrace();
- }
- }
+ ExprNodeGenericFuncDesc funcExpr = (ExprNodeGenericFuncDesc) hiveExpr;
+ GenericUDF genericUdf = funcExpr.getGenericUDF();
+ UDF udf = null;
+ if (genericUdf instanceof GenericUDFBridge) {
+ GenericUDFBridge bridge = (GenericUDFBridge) genericUdf;
+ try {
+ udf = bridge.getUdfClass().newInstance();
+ } catch (Exception e) {
+ e.printStackTrace();
+ }
+ }
- /**
- * set up the hive function
- */
- Object hiveFunction = genericUdf;
- if (udf != null)
- hiveFunction = udf;
+ /**
+ * set up the hive function
+ */
+ Object hiveFunction = genericUdf;
+ if (udf != null)
+ hiveFunction = udf;
- FunctionIdentifier funcId = HiveAlgebricksBuiltInFunctionMap.INSTANCE.getAlgebricksFunctionId(hiveFunction
- .getClass());
- if (funcId == null) {
- funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE, hiveFunction.getClass().getName());
- }
+ FunctionIdentifier funcId = HiveAlgebricksBuiltInFunctionMap.INSTANCE
+ .getAlgebricksFunctionId(hiveFunction.getClass());
+ if (funcId == null) {
+ funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE,
+ hiveFunction.getClass().getName());
+ }
- Object functionInfo = null;
- if (genericUdf instanceof GenericUDFBridge) {
- functionInfo = funcExpr;
- }
+ Object functionInfo = null;
+ if (genericUdf instanceof GenericUDFBridge) {
+ functionInfo = funcExpr;
+ }
- /**
- * generate the function call expression
- */
- ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(new HiveFunctionInfo(
- funcId, functionInfo), arguments);
- AlgebricksExpr = AlgebricksFuncExpr;
+ /**
+ * generate the function call expression
+ */
+ ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(
+ new HiveFunctionInfo(funcId, functionInfo), arguments);
+ AlgebricksExpr = AlgebricksFuncExpr;
- } else if (hiveExpr instanceof ExprNodeColumnDesc) {
- ExprNodeColumnDesc column = (ExprNodeColumnDesc) hiveExpr;
- LogicalVariable var = this.getVariable(column.getColumn());
- AlgebricksExpr = new VariableReferenceExpression(var);
+ } else if (hiveExpr instanceof ExprNodeColumnDesc) {
+ ExprNodeColumnDesc column = (ExprNodeColumnDesc) hiveExpr;
+ LogicalVariable var = this.getVariable(column.getColumn());
+ AlgebricksExpr = new VariableReferenceExpression(var);
- } else if (hiveExpr instanceof ExprNodeFieldDesc) {
- FunctionIdentifier funcId;
- funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE, ExpressionConstant.FIELDACCESS);
+ } else if (hiveExpr instanceof ExprNodeFieldDesc) {
+ FunctionIdentifier funcId;
+ funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE,
+ ExpressionConstant.FIELDACCESS);
- ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(new HiveFunctionInfo(
- funcId, hiveExpr));
- AlgebricksExpr = AlgebricksFuncExpr;
- } else if (hiveExpr instanceof ExprNodeConstantDesc) {
- ExprNodeConstantDesc hiveConst = (ExprNodeConstantDesc) hiveExpr;
- Object value = hiveConst.getValue();
- AlgebricksExpr = new ConstantExpression(new HivesterixConstantValue(value));
- } else if (hiveExpr instanceof ExprNodeNullDesc) {
- FunctionIdentifier funcId;
- funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE, ExpressionConstant.NULL);
+ ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(
+ new HiveFunctionInfo(funcId, hiveExpr));
+ AlgebricksExpr = AlgebricksFuncExpr;
+ } else if (hiveExpr instanceof ExprNodeConstantDesc) {
+ ExprNodeConstantDesc hiveConst = (ExprNodeConstantDesc) hiveExpr;
+ Object value = hiveConst.getValue();
+ AlgebricksExpr = new ConstantExpression(
+ new HivesterixConstantValue(value));
+ } else if (hiveExpr instanceof ExprNodeNullDesc) {
+ FunctionIdentifier funcId;
+ funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE,
+ ExpressionConstant.NULL);
- ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(new HiveFunctionInfo(
- funcId, hiveExpr));
+ ScalarFunctionCallExpression AlgebricksFuncExpr = new ScalarFunctionCallExpression(
+ new HiveFunctionInfo(funcId, hiveExpr));
- AlgebricksExpr = AlgebricksFuncExpr;
- } else {
- throw new IllegalStateException("unknown hive expression");
- }
- return new MutableObject<ILogicalExpression>(AlgebricksExpr);
- }
+ AlgebricksExpr = AlgebricksFuncExpr;
+ } else {
+ throw new IllegalStateException("unknown hive expression");
+ }
+ return new MutableObject<ILogicalExpression>(AlgebricksExpr);
+ }
- /**
- * translate aggregation function expression
- *
- * @param aggregateDesc
- * @return
- */
- public Mutable<ILogicalExpression> translateAggregation(AggregationDesc aggregateDesc) {
+ /**
+ * translate aggregation function expression
+ *
+ * @param aggregateDesc
+ * @return
+ */
+ public Mutable<ILogicalExpression> translateAggregation(
+ AggregationDesc aggregateDesc) {
- String UDAFName = aggregateDesc.getGenericUDAFName();
+ String UDAFName = aggregateDesc.getGenericUDAFName();
- List<Mutable<ILogicalExpression>> arguments = new ArrayList<Mutable<ILogicalExpression>>();
- List<ExprNodeDesc> children = aggregateDesc.getParameters();
+ List<Mutable<ILogicalExpression>> arguments = new ArrayList<Mutable<ILogicalExpression>>();
+ List<ExprNodeDesc> children = aggregateDesc.getParameters();
- for (ExprNodeDesc child : children)
- arguments.add(translateScalarFucntion(child));
+ for (ExprNodeDesc child : children)
+ arguments.add(translateScalarFucntion(child));
- FunctionIdentifier funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE, UDAFName + "("
- + aggregateDesc.getMode() + ")");
- HiveFunctionInfo funcInfo = new HiveFunctionInfo(funcId, aggregateDesc);
- AggregateFunctionCallExpression aggregationExpression = new AggregateFunctionCallExpression(funcInfo, false,
- arguments);
- return new MutableObject<ILogicalExpression>(aggregationExpression);
- }
+ FunctionIdentifier funcId = new FunctionIdentifier(
+ ExpressionConstant.NAMESPACE, UDAFName + "("
+ + aggregateDesc.getMode() + ")");
+ HiveFunctionInfo funcInfo = new HiveFunctionInfo(funcId, aggregateDesc);
+ AggregateFunctionCallExpression aggregationExpression = new AggregateFunctionCallExpression(
+ funcInfo, false, arguments);
+ return new MutableObject<ILogicalExpression>(aggregationExpression);
+ }
- /**
- * translate aggregation function expression
- *
- * @param aggregator
- * @return
- */
- public Mutable<ILogicalExpression> translateUnnestFunction(UDTFDesc udtfDesc, Mutable<ILogicalExpression> argument) {
+ /**
+ * translate aggregation function expression
+ *
+ * @param aggregator
+ * @return
+ */
+ public Mutable<ILogicalExpression> translateUnnestFunction(
+ UDTFDesc udtfDesc, Mutable<ILogicalExpression> argument) {
- String UDTFName = udtfDesc.getUDTFName();
+ String UDTFName = udtfDesc.getUDTFName();
- FunctionIdentifier funcId = new FunctionIdentifier(ExpressionConstant.NAMESPACE, UDTFName);
- UnnestingFunctionCallExpression unnestingExpression = new UnnestingFunctionCallExpression(new HiveFunctionInfo(
- funcId, udtfDesc));
- unnestingExpression.getArguments().add(argument);
- return new MutableObject<ILogicalExpression>(unnestingExpression);
- }
+ FunctionIdentifier funcId = new FunctionIdentifier(
+ ExpressionConstant.NAMESPACE, UDTFName);
+ UnnestingFunctionCallExpression unnestingExpression = new UnnestingFunctionCallExpression(
+ new HiveFunctionInfo(funcId, udtfDesc));
+ unnestingExpression.getArguments().add(argument);
+ return new MutableObject<ILogicalExpression>(unnestingExpression);
+ }
- /**
- * get typeinfo
- */
- @Override
- public TypeInfo getType(LogicalVariable var) {
- return variableToType.get(var);
- }
+ /**
+ * get typeinfo
+ */
+ @Override
+ public TypeInfo getType(LogicalVariable var) {
+ return variableToType.get(var);
+ }
- /**
- * get variable from variable name
- */
- @Override
- public LogicalVariable getVariable(String name) {
- return nameToLogicalVariableMap.get(name);
- }
+ /**
+ * get variable from variable name
+ */
+ @Override
+ public LogicalVariable getVariable(String name) {
+ return nameToLogicalVariableMap.get(name);
+ }
- @Override
- public LogicalVariable getVariableFromFieldName(String fieldName) {
- return this.getVariableOnly(fieldName);
- }
+ @Override
+ public LogicalVariable getVariableFromFieldName(String fieldName) {
+ return this.getVariableOnly(fieldName);
+ }
- /**
- * set the metadata provider
- */
- @Override
- public void setMetadataProvider(IMetadataProvider<PartitionDesc, Object> metadata) {
- this.metaData = metadata;
- }
+ /**
+ * set the metadata provider
+ */
+ @Override
+ public void setMetadataProvider(
+ IMetadataProvider<PartitionDesc, Object> metadata) {
+ this.metaData = metadata;
+ }
- /**
- * insert ReplicateOperator when necessary
- */
- private void insertReplicateOperator(List<Mutable<ILogicalOperator>> roots) {
- Map<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> childToParentsMap = new HashMap<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>>();
- buildChildToParentsMapping(roots, childToParentsMap);
- for (Entry<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> entry : childToParentsMap.entrySet()) {
- List<Mutable<ILogicalOperator>> pList = entry.getValue();
- if (pList.size() > 1) {
- ILogicalOperator rop = new ReplicateOperator(pList.size());
- Mutable<ILogicalOperator> ropRef = new MutableObject<ILogicalOperator>(rop);
- Mutable<ILogicalOperator> childRef = entry.getKey();
- rop.getInputs().add(childRef);
- for (Mutable<ILogicalOperator> parentRef : pList) {
- ILogicalOperator parentOp = parentRef.getValue();
- int index = parentOp.getInputs().indexOf(childRef);
- parentOp.getInputs().set(index, ropRef);
- }
- }
- }
- }
+ /**
+ * insert ReplicateOperator when necessary
+ */
+ private void insertReplicateOperator(List<Mutable<ILogicalOperator>> roots) {
+ Map<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> childToParentsMap = new HashMap<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>>();
+ buildChildToParentsMapping(roots, childToParentsMap);
+ for (Entry<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> entry : childToParentsMap
+ .entrySet()) {
+ List<Mutable<ILogicalOperator>> pList = entry.getValue();
+ if (pList.size() > 1) {
+ ILogicalOperator rop = new ReplicateOperator(pList.size());
+ Mutable<ILogicalOperator> ropRef = new MutableObject<ILogicalOperator>(
+ rop);
+ Mutable<ILogicalOperator> childRef = entry.getKey();
+ rop.getInputs().add(childRef);
+ for (Mutable<ILogicalOperator> parentRef : pList) {
+ ILogicalOperator parentOp = parentRef.getValue();
+ int index = parentOp.getInputs().indexOf(childRef);
+ parentOp.getInputs().set(index, ropRef);
+ }
+ }
+ }
+ }
- /**
- * build the mapping from child to Parents
- *
- * @param roots
- * @param childToParentsMap
- */
- private void buildChildToParentsMapping(List<Mutable<ILogicalOperator>> roots,
- Map<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> map) {
- for (Mutable<ILogicalOperator> opRef : roots) {
- List<Mutable<ILogicalOperator>> childRefs = opRef.getValue().getInputs();
- for (Mutable<ILogicalOperator> childRef : childRefs) {
- List<Mutable<ILogicalOperator>> parentList = map.get(childRef);
- if (parentList == null) {
- parentList = new ArrayList<Mutable<ILogicalOperator>>();
- map.put(childRef, parentList);
- }
- if (!parentList.contains(opRef))
- parentList.add(opRef);
- }
- buildChildToParentsMapping(childRefs, map);
- }
- }
+ /**
+ * build the mapping from child to Parents
+ *
+ * @param roots
+ * @param childToParentsMap
+ */
+ private void buildChildToParentsMapping(
+ List<Mutable<ILogicalOperator>> roots,
+ Map<Mutable<ILogicalOperator>, List<Mutable<ILogicalOperator>>> map) {
+ for (Mutable<ILogicalOperator> opRef : roots) {
+ List<Mutable<ILogicalOperator>> childRefs = opRef.getValue()
+ .getInputs();
+ for (Mutable<ILogicalOperator> childRef : childRefs) {
+ List<Mutable<ILogicalOperator>> parentList = map.get(childRef);
+ if (parentList == null) {
+ parentList = new ArrayList<Mutable<ILogicalOperator>>();
+ map.put(childRef, parentList);
+ }
+ if (!parentList.contains(opRef))
+ parentList.add(opRef);
+ }
+ buildChildToParentsMapping(childRefs, map);
+ }
+ }
}
diff --git a/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/exec/HyracksExecutionEngine.java b/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/exec/HyracksExecutionEngine.java
index 6c1ac72..ed5ab70 100644
--- a/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/exec/HyracksExecutionEngine.java
+++ b/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/exec/HyracksExecutionEngine.java
@@ -74,8 +74,6 @@
private static final Log LOG = LogFactory.getLog(HyracksExecutionEngine.class.getName());
- // private static final String[] locConstraints = {}
-
private static List<Pair<AbstractRuleController, List<IAlgebraicRewriteRule>>> DEFAULT_LOGICAL_REWRITES = new ArrayList<Pair<AbstractRuleController, List<IAlgebraicRewriteRule>>>();
private static List<Pair<AbstractRuleController, List<IAlgebraicRewriteRule>>> DEFAULT_PHYSICAL_REWRITES = new ArrayList<Pair<AbstractRuleController, List<IAlgebraicRewriteRule>>>();
static {
@@ -273,7 +271,6 @@
* list of map-reduce tasks
*/
Task<? extends Serializable> task = rootTasks.get(i);
- // System.out.println("!" + task.getName());
if (task instanceof MapRedTask) {
List<Operator> mapRootOps = articulateMapReduceOperators(task, rootOps, aliasToPath, rootTasks);
diff --git a/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/operator/filescan/HiveFileScanOperatorDescriptor.java b/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/operator/filescan/HiveFileScanOperatorDescriptor.java
index d248486..9e62c73 100644
--- a/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/operator/filescan/HiveFileScanOperatorDescriptor.java
+++ b/hivesterix/src/main/java/edu/uci/ics/hivesterix/runtime/operator/filescan/HiveFileScanOperatorDescriptor.java
@@ -110,7 +110,7 @@
if (returnedSplits[i] != null)
splits[i] = returnedSplits[i];
- System.out.println("!!! number of splits: " + splits.length);
+ System.out.println("number of splits: " + splits.length);
constraintsByHostNames = new String[splits.length];
for (int i = 0; i < splits.length; i++) {
try {