From a3b33db888808495d631a623a1c71c25fba3954d Mon Sep 17 00:00:00 2001
From: Volodymyr Vysotskyi <vvovyk@gmail.com>
Date: Thu, 1 Nov 2018 18:31:16 +0200
Subject: [PATCH] DRILL-4456: Add Hive translate UDF

closes #1527
---
 .../exec/expr/fn/HiveFunctionRegistry.java    | 45 ++++++++++++-------
 .../exec/fn/hive/TestInbuiltHiveUDFs.java     | 13 ++++++
 2 files changed, 43 insertions(+), 15 deletions(-)

diff --git a/contrib/storage-hive/core/src/main/java/org/apache/drill/exec/expr/fn/HiveFunctionRegistry.java b/contrib/storage-hive/core/src/main/java/org/apache/drill/exec/expr/fn/HiveFunctionRegistry.java
index cb00ede9bf0..0ec97785372 100644
--- a/contrib/storage-hive/core/src/main/java/org/apache/drill/exec/expr/fn/HiveFunctionRegistry.java
+++ b/contrib/storage-hive/core/src/main/java/org/apache/drill/exec/expr/fn/HiveFunctionRegistry.java
@@ -20,9 +20,12 @@
 import java.util.HashSet;
 import java.util.Map;
 import java.util.Set;
+import java.util.stream.Stream;
 
 import org.apache.calcite.rel.type.RelDataType;
 import org.apache.calcite.sql.SqlOperatorBinding;
+import org.apache.calcite.sql.fun.OracleSqlOperatorTable;
+import org.apache.calcite.sql.fun.SqlStdOperatorTable;
 import org.apache.calcite.sql.type.SqlReturnTypeInference;
 import org.apache.calcite.sql.type.SqlTypeName;
 import org.apache.drill.common.config.DrillConfig;
@@ -38,6 +41,7 @@
 import org.apache.drill.exec.planner.sql.HiveUDFOperator;
 import org.apache.drill.exec.planner.sql.HiveUDFOperatorWithoutInference;
 import org.apache.drill.exec.planner.sql.TypeInferenceUtils;
+import org.apache.drill.shaded.guava.com.google.common.collect.ImmutableMap;
 import org.apache.hadoop.hive.ql.exec.Description;
 import org.apache.hadoop.hive.ql.exec.UDF;
 import org.apache.hadoop.hive.ql.udf.UDFType;
@@ -48,8 +52,18 @@
 import org.apache.drill.shaded.guava.com.google.common.collect.ArrayListMultimap;
 import org.apache.drill.shaded.guava.com.google.common.collect.Sets;
 
-public class HiveFunctionRegistry implements PluggableFunctionRegistry{
-  static final org.slf4j.Logger logger = org.slf4j.LoggerFactory.getLogger(HiveFunctionRegistry.class);
+public class HiveFunctionRegistry implements PluggableFunctionRegistry {
+  private static final org.slf4j.Logger logger = org.slf4j.LoggerFactory.getLogger(HiveFunctionRegistry.class);
+
+  /**
+   * Map for renaming UDFs. Keys of the map represent UDF names which should be replaced
+   * and its values represent target UDF names.
+   */
+  private static final Map<String, String> FUNCTION_REPLACE_MAP = ImmutableMap.<String, String> builder()
+      // renames Hive's TRANSLATE UDF to TRANSLATE3 due to CALCITE-1115
+      .put(SqlStdOperatorTable.TRANSLATE.getName().toLowerCase(),
+          OracleSqlOperatorTable.TRANSLATE3.getName().toLowerCase())
+      .build();
 
   private ArrayListMultimap<String, Class<? extends GenericUDF>> methodsGenericUDF = ArrayListMultimap.create();
   private ArrayListMultimap<String, Class<? extends UDF>> methodsUDF = ArrayListMultimap.create();
@@ -102,27 +116,28 @@ public void register(DrillOperatorTable operatorTable) {
     }
   }
 
-  private <C,I> void register(Class<? extends I> clazz, ArrayListMultimap<String,Class<? extends I>> methods) {
+  private <I> void register(Class<? extends I> clazz, ArrayListMultimap<String, Class<? extends I>> methods) {
     Description desc = clazz.getAnnotation(Description.class);
-    String[] names;
+    Stream<String> namesStream;
     if (desc != null) {
-      names = desc.name().split(",");
-      for (int i=0; i<names.length; i++) {
-        names[i] = names[i].trim();
-      }
-    }else{
-      names = new String[]{clazz.getName().replace('.', '_')};
+      namesStream = Stream.of(desc.name().split(","))
+          .map(String::trim);
+    } else {
+      namesStream = Stream.of(clazz)
+          .map(Class::getName)
+          .map(name -> name.replace('.', '_'));
     }
 
+    // Checks specified array of function names whether they should be replaced
+    // using FUNCTION_REPLACE_MAP map.
+    namesStream.map(String::toLowerCase)
+        .map(functionName -> FUNCTION_REPLACE_MAP.getOrDefault(functionName, functionName))
+        .forEach(name -> methods.put(name, clazz));
+
     UDFType type = clazz.getAnnotation(UDFType.class);
     if (type != null && !type.deterministic()) {
       nonDeterministicUDFs.add(clazz);
     }
-
-
-    for(int i=0; i<names.length;i++) {
-      methods.put(names[i].toLowerCase(), clazz);
-    }
   }
 
   /**
diff --git a/contrib/storage-hive/core/src/test/java/org/apache/drill/exec/fn/hive/TestInbuiltHiveUDFs.java b/contrib/storage-hive/core/src/test/java/org/apache/drill/exec/fn/hive/TestInbuiltHiveUDFs.java
index c74a7c40316..43c3e3f9aea 100644
--- a/contrib/storage-hive/core/src/test/java/org/apache/drill/exec/fn/hive/TestInbuiltHiveUDFs.java
+++ b/contrib/storage-hive/core/src/test/java/org/apache/drill/exec/fn/hive/TestInbuiltHiveUDFs.java
@@ -184,4 +184,17 @@ public void testToUTCTimestamp() throws Exception {
         .go();
   }
 
+  @Test // DRILL-4456
+  public void testTranslate3() throws Exception {
+    testBuilder()
+        .sqlQuery("SELECT translate(string_field, 's', 'S') as ts," +
+            "translate(varchar_field, 'v', 'V') as tv,\n" +
+            "translate('literal', 'l', 'L') as tl from hive.readtest")
+        .unOrdered()
+        .baselineColumns("ts", "tv", "tl")
+        .baselineValues("Stringfield", "Varcharfield", "LiteraL")
+        .baselineValues(null, null, "LiteraL")
+        .go();
+  }
+
 }