[sidecar] Fix ANALYZE queries on varchar columns when sidecar is enabled

Author: pdabre12

presto-main-base/src/main/java/com/facebook/presto/sql/planner/StatisticsAggregationPlanner.java

@@ -43,8 +43,10 @@
 import java.util.Optional;
 import java.util.stream.Collectors;
 
+import static com.facebook.presto.SystemSessionProperties.isNativeExecutionEnabled;
 import static com.facebook.presto.SystemSessionProperties.shouldOptimizerUseHistograms;
 import static com.facebook.presto.common.type.BigintType.BIGINT;
+import static com.facebook.presto.common.type.StandardTypes.VARCHAR;
 import static com.facebook.presto.common.type.UnknownType.UNKNOWN;
 import static com.facebook.presto.spi.StandardErrorCode.NOT_SUPPORTED;
 import static com.facebook.presto.spi.function.FunctionKind.AGGREGATE;
@@ -180,7 +182,17 @@ private ColumnStatisticsAggregation createColumnAggregationFromFunctionName(Colu
         FunctionMetadata functionMeta = functionAndTypeResolver.getFunctionMetadata(functionHandle);
         Type inputType = functionAndTypeResolver.getType(getOnlyElement(functionMeta.getArgumentTypes()));
         Type outputType = functionAndTypeResolver.getType(functionMeta.getReturnType());
-        verify(inputType.equals(input.getType()) || input.getType().equals(UNKNOWN), "resolved function input type does not match the input type: %s != %s", inputType, input.getType());
+
+        // todo: fix this hack
+        // In native clusters, we do not support parameterized varchar,
+        // hence if its a varchar type, we just compare the type signature base.
+        boolean isVarcharType = input.getType().getTypeSignature().getBase().equals(VARCHAR);
+        boolean isTypeSignatureBaseMatching = inputType.getTypeSignature().getBase().equals(input.getType().getTypeSignature().getBase());
+        verify(
+                inputType.equals(input.getType()) ||
+                        input.getType().equals(UNKNOWN) ||
+                        isNativeExecutionEnabled(session) && isVarcharType && isTypeSignatureBaseMatching,
+                "resolved function input type does not match the input type: %s != %s", inputType, input.getType());
         return new ColumnStatisticsAggregation(
                 new AggregationNode.Aggregation(
                         new CallExpression(

presto-native-sidecar-plugin/src/test/java/com/facebook/presto/sidecar/TestNativeSidecarPlugin.java

@@ -33,6 +33,7 @@
 import static com.facebook.presto.nativeworker.NativeQueryRunnerUtils.createNation;
 import static com.facebook.presto.nativeworker.NativeQueryRunnerUtils.createOrders;
 import static com.facebook.presto.nativeworker.NativeQueryRunnerUtils.createOrdersEx;
+import static com.facebook.presto.nativeworker.NativeQueryRunnerUtils.createRegion;
 import static com.facebook.presto.sidecar.NativeSidecarPluginQueryRunnerUtils.setupNativeSidecarPlugin;
 import static java.lang.String.format;
 import static org.testng.Assert.assertEquals;
@@ -54,6 +55,7 @@ protected void createTables()
         createNation(queryRunner);
         createOrders(queryRunner);
         createOrdersEx(queryRunner);
+        createRegion(queryRunner);
     }
 
     @Override
@@ -234,6 +236,49 @@ public void testShowStats()
         }
     }
 
+    @Test
+    public void testAnalyzeStats()
+    {
+        assertUpdate("ANALYZE region", 5);
+
+        // Show stats returns the following stats for each column in region table:
+        // column_name | data_size | distinct_values_count | nulls_fraction | row_count | low_value | high_value
+        assertQuery("SHOW STATS FOR region",
+                "SELECT * FROM (VALUES" +
+                        "('regionkey', NULL, 5e0, 0e0, NULL, '0', '4', NULL)," +
+                        "('name', 5.4e1, 5e0, 0e0, NULL, NULL, NULL, NULL)," +
+                        "('comment', 3.5e2, 5e0, 0e0, NULL, NULL, NULL, NULL)," +
+                        "(NULL, NULL, NULL, NULL, 5e0, NULL, NULL, NULL))");
+
+        // Create a partitioned table and run analyze on it.
+        String tmpTableName = generateRandomTableName();
+        try {
+            getQueryRunner().execute(String.format("CREATE TABLE %s (name VARCHAR, regionkey BIGINT," +
+                    "nationkey BIGINT) WITH (partitioned_by = ARRAY['regionkey','nationkey'])", tmpTableName));
+            getQueryRunner().execute(
+                    String.format("INSERT INTO %s SELECT name, regionkey, nationkey FROM nation", tmpTableName));
+            assertQuery(String.format("SELECT * FROM %s", tmpTableName),
+                    "SELECT name, regionkey, nationkey FROM nation");
+            assertUpdate(String.format("ANALYZE %s", tmpTableName), 25);
+            assertQuery(String.format("SHOW STATS for %s", tmpTableName),
+                    "SELECT * FROM (VALUES" +
+                            "('name', 2.77e2, 1e0, 0e0, NULL, NULL, NULL, NULL)," +
+                            "('regionkey', NULL, 5e0, 0e0, NULL, '0', '4', NULL)," +
+                            "('nationkey', NULL, 2.5e1, 0e0, NULL, '0', '24', NULL)," +
+                            "(NULL, NULL, NULL, NULL, 2.5e1, NULL, NULL, NULL))");
+            assertUpdate(String.format("ANALYZE %s WITH (partitions = ARRAY[ARRAY['0','0'],ARRAY['4', '11']])", tmpTableName), 2);
+            assertQuery(String.format("SHOW STATS for (SELECT * FROM %s where regionkey=4 and nationkey=11)", tmpTableName),
+                    "SELECT * FROM (VALUES" +
+                            "('name', 8e0, 1e0, 0e0, NULL, NULL, NULL, NULL)," +
+                            "('regionkey', NULL, 1e0, 0e0, NULL, '4', '4', NULL)," +
+                            "('nationkey', NULL, 1e0, 0e0, NULL, '11', '11', NULL)," +
+                            "(NULL, NULL, NULL, NULL, 1e0, NULL, NULL, NULL))");
+        }
+        finally {
+            dropTableIfExists(tmpTableName);
+        }
+    }
+
     private String generateRandomTableName()
     {
         String tableName = "tmp_presto_" + UUID.randomUUID().toString().replace("-", "");