Fix failure when matching empty dictionary

lib/trino-parquet/src/main/java/io/trino/parquet/predicate/TupleDomainParquetPredicate.java

@@ -485,6 +485,11 @@ private static Domain getDomain(Type type, DictionaryDescriptor dictionaryDescri
         }
 
         int dictionarySize = dictionaryPage.get().getDictionarySize();
+
+        if (dictionarySize == 0) {
+            return Domain.onlyNull(type);
+        }
+
         DictionaryValueConverter converter = new DictionaryValueConverter(dictionary);
         Function<Integer, Object> convertFunction = converter.getConverter(columnDescriptor.getPrimitiveType());
         List<Object> values = new ArrayList<>();

lib/trino-parquet/src/test/java/io/trino/parquet/TestTupleDomainParquetPredicate.java

@@ -69,6 +69,7 @@
 import static io.trino.spi.predicate.Domain.all;
 import static io.trino.spi.predicate.Domain.create;
 import static io.trino.spi.predicate.Domain.notNull;
+import static io.trino.spi.predicate.Domain.onlyNull;
 import static io.trino.spi.predicate.Domain.singleValue;
 import static io.trino.spi.predicate.Range.range;
 import static io.trino.spi.predicate.TupleDomain.withColumnDomains;
@@ -612,6 +613,39 @@ public void testVarcharMatchesWithDictionaryDescriptor()
         assertTrue(parquetPredicate.matches(new DictionaryDescriptor(column, Optional.of(page))));
     }
 
+    @Test
+    public void testEmptyDictionary()
+    {
+        ColumnDescriptor columnDescriptor = new ColumnDescriptor(new String[] {"path"}, Types.optional(BINARY).named("Test column"), 0, 0);
+        RichColumnDescriptor column = new RichColumnDescriptor(columnDescriptor, new PrimitiveType(OPTIONAL, BINARY, "Test column"));
+        ColumnDescriptor descriptor = new ColumnDescriptor(column.getPath(), column.getPrimitiveType(), 0, 0);
+        VarcharType type = createVarcharType(255);
+
+        DictionaryPage dictionary = new DictionaryPage(EMPTY_SLICE, 0, PLAIN_DICTIONARY);
+        TupleDomainParquetPredicate predicate;
+
+        // only non-nulls allowed
+        predicate = new TupleDomainParquetPredicate(
+                withColumnDomains(singletonMap(descriptor, notNull(type))),
+                singletonList(column),
+                UTC);
+        assertFalse(predicate.matches(new DictionaryDescriptor(column, Optional.of(dictionary))));
+
+        // only nulls allowed
+        predicate = new TupleDomainParquetPredicate(
+                withColumnDomains(singletonMap(descriptor, onlyNull(type))),
+                singletonList(column),
+                UTC);
+        assertTrue(predicate.matches(new DictionaryDescriptor(column, Optional.of(dictionary))));
+
+        // mixed non-nulls and nulls allowed
+        predicate = new TupleDomainParquetPredicate(
+                withColumnDomains(singletonMap(descriptor, singleValue(type, EMPTY_SLICE, true))),
+                singletonList(column),
+                UTC);
+        assertTrue(predicate.matches(new DictionaryDescriptor(column, Optional.of(dictionary))));
+    }
+
     @Test
     public void testColumnIndexWithNullPages()
             throws Exception

plugin/trino-hive/src/test/java/io/trino/plugin/hive/parquet/TestOnlyNulls.java

@@ -0,0 +1,107 @@
+/*
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package io.trino.plugin.hive.parquet;
+
+import com.google.common.io.Resources;
+import io.trino.plugin.hive.HiveColumnHandle;
+import io.trino.plugin.hive.HiveConfig;
+import io.trino.plugin.hive.HivePageSourceFactory;
+import io.trino.plugin.hive.HiveStorageFormat;
+import io.trino.plugin.hive.HiveType;
+import io.trino.plugin.hive.acid.AcidTransaction;
+import io.trino.plugin.hive.benchmark.StandardFileFormats;
+import io.trino.spi.connector.ConnectorPageSource;
+import io.trino.spi.predicate.Domain;
+import io.trino.spi.predicate.TupleDomain;
+import io.trino.spi.type.Type;
+import io.trino.testing.MaterializedResult;
+import io.trino.testing.MaterializedRow;
+import org.apache.hadoop.fs.Path;
+import org.testng.annotations.Test;
+
+import java.io.File;
+import java.util.List;
+import java.util.Map;
+import java.util.Optional;
+import java.util.OptionalInt;
+import java.util.Properties;
+
+import static io.trino.hadoop.ConfigurationInstantiator.newEmptyConfiguration;
+import static io.trino.plugin.hive.HiveColumnHandle.ColumnType.REGULAR;
+import static io.trino.plugin.hive.HiveColumnHandle.createBaseColumn;
+import static io.trino.plugin.hive.HiveTestUtils.HDFS_ENVIRONMENT;
+import static io.trino.plugin.hive.HiveTestUtils.getHiveSession;
+import static io.trino.spi.type.IntegerType.INTEGER;
+import static io.trino.testing.MaterializedResult.materializeSourceDataStream;
+import static java.util.Collections.singletonList;
+import static org.apache.hadoop.hive.serde.serdeConstants.SERIALIZATION_LIB;
+import static org.assertj.core.api.Assertions.assertThat;
+
+public class TestOnlyNulls
+{
+    @Test
+    public void testOnlyNulls()
+            throws Exception
+    {
+        // The file contains only nulls and a dictionary page with 0 entries.
+        File parquetFile = new File(Resources.getResource("issue-10873.parquet").toURI());
+        String columnName = "x";
+        Type columnType = INTEGER;
+
+        HiveColumnHandle column = createBaseColumn(columnName, 0, HiveType.toHiveType(columnType), columnType, REGULAR, Optional.empty());
+
+        // match not null
+        try (ConnectorPageSource pageSource = createPageSource(parquetFile, column, TupleDomain.withColumnDomains(Map.of(column, Domain.notNull(columnType))))) {
+            MaterializedResult result = materializeSourceDataStream(getHiveSession(new HiveConfig()), pageSource, List.of(columnType)).toTestTypes();
+            assertThat(result.getMaterializedRows()).isEmpty();
+        }
+
+        // match null
+        try (ConnectorPageSource pageSource = createPageSource(parquetFile, column, TupleDomain.withColumnDomains(Map.of(column, Domain.onlyNull(columnType))))) {
+            MaterializedResult result = materializeSourceDataStream(getHiveSession(new HiveConfig()), pageSource, List.of(columnType)).toTestTypes();
+
+            assertThat(result.getMaterializedRows())
+                    .isEqualTo(List.of(
+                            new MaterializedRow(singletonList(null)),
+                            new MaterializedRow(singletonList(null)),
+                            new MaterializedRow(singletonList(null)),
+                            new MaterializedRow(singletonList(null))));
+        }
+    }
+
+    private static ConnectorPageSource createPageSource(File parquetFile, HiveColumnHandle column, TupleDomain<HiveColumnHandle> domain)
+    {
+        HivePageSourceFactory pageSourceFactory = StandardFileFormats.TRINO_PARQUET.getHivePageSourceFactory(HDFS_ENVIRONMENT).orElseThrow();
+
+        Properties schema = new Properties();
+        schema.setProperty(SERIALIZATION_LIB, HiveStorageFormat.PARQUET.getSerde());
+
+        return pageSourceFactory.createPageSource(
+                        newEmptyConfiguration(),
+                        getHiveSession(new HiveConfig()),
+                        new Path(parquetFile.toURI()),
+                        0,
+                        parquetFile.length(),
+                        parquetFile.length(),
+                        schema,
+                        List.of(column),
+                        domain,
+                        Optional.empty(),
+                        OptionalInt.empty(),
+                        false,
+                        AcidTransaction.NO_ACID_TRANSACTION)
+                .orElseThrow()
+                .get();
+    }
+}