[HUDI-2385] Make parquet dictionary encoding configurable (#3578)

Co-authored-by: leesf <leesf@apache.org>

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieStorageConfig.java

@@ -110,6 +110,11 @@ public class HoodieStorageConfig extends HoodieConfig {
       .defaultValue("gzip")
       .withDocumentation("Compression Codec for parquet files");
 
+  public static final ConfigProperty<Boolean> PARQUET_DICTIONARY_ENABLED = ConfigProperty
+      .key("hoodie.parquet.dictionary.enabled")
+      .defaultValue(true)
+      .withDocumentation("Whether to use dictionary encoding");
+
   public static final ConfigProperty<String> HFILE_COMPRESSION_ALGORITHM_NAME = ConfigProperty
       .key("hoodie.hfile.compression.algorithm")
       .defaultValue("GZ")

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieWriteConfig.java

@@ -1403,10 +1403,6 @@ public class HoodieWriteConfig extends HoodieConfig {
     return getInt(HoodieStorageConfig.LOGFILE_DATA_BLOCK_MAX_SIZE);
   }
 
-  public long getLogFileMaxSize() {
-    return getLong(HoodieStorageConfig.LOGFILE_MAX_SIZE);
-  }
-
   public double getParquetCompressionRatio() {
     return getDouble(HoodieStorageConfig.PARQUET_COMPRESSION_RATIO_FRACTION);
   }
@@ -1415,6 +1411,14 @@ public class HoodieWriteConfig extends HoodieConfig {
     return CompressionCodecName.fromConf(getString(HoodieStorageConfig.PARQUET_COMPRESSION_CODEC_NAME));
   }
 
+  public boolean parquetDictionaryEnabled() {
+    return getBoolean(HoodieStorageConfig.PARQUET_DICTIONARY_ENABLED);
+  }
+
+  public long getLogFileMaxSize() {
+    return getLong(HoodieStorageConfig.LOGFILE_MAX_SIZE);
+  }
+
   public double getLogFileToParquetCompressionRatio() {
     return getDouble(HoodieStorageConfig.LOGFILE_TO_PARQUET_COMPRESSION_RATIO_FRACTION);
   }

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/io/storage/HoodieAvroParquetConfig.java

@@ -33,4 +33,10 @@ public class HoodieAvroParquetConfig extends HoodieBaseParquetConfig<HoodieAvroW
                                  double compressionRatio) {
     super(writeSupport, compressionCodecName, blockSize, pageSize, maxFileSize, hadoopConf, compressionRatio);
   }
+
+  public HoodieAvroParquetConfig(HoodieAvroWriteSupport writeSupport, CompressionCodecName compressionCodecName,
+      int blockSize, int pageSize, long maxFileSize, Configuration hadoopConf,
+      double compressionRatio, boolean directoryEnabled) {
+    super(writeSupport, compressionCodecName, blockSize, pageSize, maxFileSize, hadoopConf, compressionRatio, directoryEnabled);
+  }
 }

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/io/storage/HoodieBaseParquetConfig.java

@@ -27,15 +27,21 @@ import org.apache.parquet.hadoop.metadata.CompressionCodecName;
  */
 public class HoodieBaseParquetConfig<T> {
   private final T writeSupport;
-  private CompressionCodecName compressionCodecName;
+  private final CompressionCodecName compressionCodecName;
-  private int blockSize;
+  private final int blockSize;
-  private int pageSize;
+  private final int pageSize;
-  private long maxFileSize;
+  private final long maxFileSize;
-  private Configuration hadoopConf;
+  private final Configuration hadoopConf;
-  private double compressionRatio;
+  private final double compressionRatio;
+  private final boolean dictionaryEnabled;
 
   public HoodieBaseParquetConfig(T writeSupport, CompressionCodecName compressionCodecName, int blockSize,
       int pageSize, long maxFileSize, Configuration hadoopConf, double compressionRatio) {
+    this(writeSupport, compressionCodecName, blockSize, pageSize, maxFileSize, hadoopConf, compressionRatio, false);
+  }
+
+  public HoodieBaseParquetConfig(T writeSupport, CompressionCodecName compressionCodecName, int blockSize,
+                                 int pageSize, long maxFileSize, Configuration hadoopConf, double compressionRatio, boolean dictionaryEnabled) {
     this.writeSupport = writeSupport;
     this.compressionCodecName = compressionCodecName;
     this.blockSize = blockSize;
@@ -43,6 +49,7 @@ public class HoodieBaseParquetConfig<T> {
     this.maxFileSize = maxFileSize;
     this.hadoopConf = hadoopConf;
     this.compressionRatio = compressionRatio;
+    this.dictionaryEnabled = dictionaryEnabled;
   }
 
   public CompressionCodecName getCompressionCodecName() {
@@ -72,4 +79,8 @@ public class HoodieBaseParquetConfig<T> {
   public T getWriteSupport() {
     return writeSupport;
   }
+
+  public boolean dictionaryEnabled() {
+    return dictionaryEnabled;
+  }
 }

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/io/storage/HoodieFileWriterFactory.java

@@ -71,7 +71,7 @@ public class HoodieFileWriterFactory {
 
     HoodieAvroParquetConfig parquetConfig = new HoodieAvroParquetConfig(writeSupport, config.getParquetCompressionCodec(),
         config.getParquetBlockSize(), config.getParquetPageSize(), config.getParquetMaxFileSize(),
-        hoodieTable.getHadoopConf(), config.getParquetCompressionRatio());
+        hoodieTable.getHadoopConf(), config.getParquetCompressionRatio(), config.parquetDictionaryEnabled());
 
     return new HoodieParquetWriter<>(instantTime, path, parquetConfig, schema, taskContextSupplier, populateMetaFields);
   }

hudi-client/hudi-client-common/src/main/java/org/apache/hudi/io/storage/HoodieParquetWriter.java

@@ -56,7 +56,7 @@ public class HoodieParquetWriter<T extends HoodieRecordPayload, R extends Indexe
     super(HoodieWrapperFileSystem.convertToHoodiePath(file, parquetConfig.getHadoopConf()),
         ParquetFileWriter.Mode.CREATE, parquetConfig.getWriteSupport(), parquetConfig.getCompressionCodecName(),
         parquetConfig.getBlockSize(), parquetConfig.getPageSize(), parquetConfig.getPageSize(),
-        DEFAULT_IS_DICTIONARY_ENABLED, DEFAULT_IS_VALIDATING_ENABLED,
+        parquetConfig.dictionaryEnabled(), DEFAULT_IS_VALIDATING_ENABLED,
         DEFAULT_WRITER_VERSION, FSUtils.registerFileSystem(file, parquetConfig.getHadoopConf()));
     this.file = HoodieWrapperFileSystem.convertToHoodiePath(file, parquetConfig.getHadoopConf());
     this.fs =