Package parquet.schema

Examples of parquet.schema.MessageType

72
73
74
75
76
77
78
79
                        mapping.source.getType(),
                        mapping.source.getPath()[0]));
                properties.add(mapping.target);
            }
        }
        this.materializeSchema = new MessageType(schema.getName(), fields);
        this.root = new DataModelConverter(properties);
    }
View Full Code Here
87
88
89
90
91
92
93
94
95
        List<Type> fields = new ArrayList<Type>();
        for (PropertyDescriptor property : descriptor.getPropertyDescriptors()) {
            Type field = computeParquetType(property);
            fields.add(field);
        }
        return new MessageType(
                descriptor.getDataModelClass().getName(),
                fields);
    }
View Full Code Here
53
54
55
56
57
58
59
60
61
  }

  private void testConversion(final String columnNamesStr, final String columnsTypeStr, final String expectedSchema) throws Exception {
    final List<String> columnNames = createHiveColumnsFrom(columnNamesStr);
    final List<TypeInfo> columnTypes = createHiveTypeInfoFrom(columnsTypeStr);
    final MessageType messageTypeFound = HiveSchemaConverter.convert(columnNames, columnTypes);
    final MessageType expectedMT = MessageTypeParser.parseMessageType(expectedSchema);
    assertEquals("converting " + columnNamesStr + ": " + columnsTypeStr + " to " + expectedSchema, expectedMT, messageTypeFound);
  }
View Full Code Here
32
33
34
35
36
37
38
39
import parquet.schema.Type.Repetition;

public class HiveSchemaConverter {

  public static MessageType convert(final List<String> columnNames, final List<TypeInfo> columnTypes) {
    final MessageType schema = new MessageType("hive_schema", convertTypes(columnNames, columnTypes));
    return schema;
  }
View Full Code Here
101
102
103
104
105
106
107
108
109
110
  }

  private void testTajoToParquetConversion(
      Schema tajoSchema, String schemaString) throws Exception {
    TajoSchemaConverter converter = new TajoSchemaConverter();
    MessageType schema = converter.convert(tajoSchema);
    MessageType expected = MessageTypeParser.parseMessageType(schemaString);
    assertEquals("converting " + schema + " to " + schemaString,
                 expected.toString(), schema.toString());
  }
View Full Code Here
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
    // remove the last semicolon, java really needs a join method for strings...
    // TODO - nvm apparently it requires a semicolon after every field decl, might want to file a bug
    // messageSchema = messageSchema.substring(schemaType, messageSchema.length() - 1);
    messageSchema += "}";

    MessageType schema = MessageTypeParser.parseMessageType(messageSchema);

    CompressionCodecName codec = CompressionCodecName.UNCOMPRESSED;
    ParquetFileWriter w = new ParquetFileWriter(configuration, schema, path);
    w.start();
    HashMap<String, Integer> columnValuesWritten = new HashMap();
    int valsWritten;
    for (int k = 0; k < numberRowGroups; k++) {
      w.startBlock(1);
      currentBooleanByte = 0;
      booleanBitCounter.reset();

      for (FieldInfo fieldInfo : fields.values()) {

        if (!columnValuesWritten.containsKey(fieldInfo.name)) {
          columnValuesWritten.put((String) fieldInfo.name, 0);
          valsWritten = 0;
        } else {
          valsWritten = columnValuesWritten.get(fieldInfo.name);
        }

        String[] path1 = { (String) fieldInfo.name };
        ColumnDescriptor c1 = schema.getColumnDescription(path1);

        w.startColumn(c1, recordsPerRowGroup, codec);
        int valsPerPage = (int) Math.ceil(recordsPerRowGroup / (float) fieldInfo.numberOfPages);
        byte[] bytes;
        // for variable length binary fields
View Full Code Here
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
    // remove the last semicolon, java really needs a join method for strings...
    // TODO - nvm apparently it requires a semicolon after every field decl, might want to file a bug
    //messageSchema = messageSchema.substring(schemaType, messageSchema.length() - 1);
    messageSchema += "}";

    MessageType schema = MessageTypeParser.parseMessageType(messageSchema);

    CompressionCodecName codec = CompressionCodecName.UNCOMPRESSED;
    ParquetFileWriter w = new ParquetFileWriter(configuration, schema, path);
    w.start();
    HashMap<String, Integer> columnValuesWritten = new HashMap();
    int valsWritten;
    for (int k = 0; k < props.numberRowGroups; k++){
      w.startBlock(1);
      currentBooleanByte = 0;
      booleanBitCounter.reset();

      for (FieldInfo fieldInfo : props.fields.values()) {

        if ( ! columnValuesWritten.containsKey(fieldInfo.name)){
          columnValuesWritten.put((String) fieldInfo.name, 0);
          valsWritten = 0;
        } else {
          valsWritten = columnValuesWritten.get(fieldInfo.name);
        }

        String[] path1 = {(String) fieldInfo.name};
        ColumnDescriptor c1 = schema.getColumnDescription(path1);

        w.startColumn(c1, props.recordsPerRowGroup, codec);
        int valsPerPage = (int) Math.ceil(props.recordsPerRowGroup / (float) fieldInfo.numberOfPages);
        byte[] bytes;
        // for variable length binary fields
View Full Code Here
78
79
80
81
82
83
84
85
86
87
  @Override
  public ReadContext init(InitContext context) {
    if (requestedSchema == null) {
      throw new RuntimeException("requestedSchema is null.");
    }
    MessageType requestedParquetSchema =
      new TajoSchemaConverter().convert(requestedSchema);
    LOG.debug("Reading data with projection:\n" + requestedParquetSchema);
    return new ReadContext(requestedParquetSchema);
  }
View Full Code Here
106
107
108
109
110
111
112
113
114
      tajoReadSchema = CatalogGsonHelper.fromJson(
          metadataReadSchema, Schema.class);
    } else {
      tajoReadSchema = readSchema;
    }
    MessageType parquetRequestedSchema = readContext.getRequestedSchema();
    return new TajoRecordMaterializer(parquetRequestedSchema, requestedSchema,
                                      tajoReadSchema);
  }
View Full Code Here
147
148
149
150
151
152
153
      if (column.getDataType().getType() == TajoDataTypes.Type.NULL_TYPE) {
        continue;
      }
      types.add(convertColumn(column));
    }
    return new MessageType(TABLE_SCHEMA, types);
  }
View Full Code Here
TOP

Related Classes of parquet.schema.MessageType

  • com.asakusafw.directio.hive.parquet.DataModelMaterializer
  • com.asakusafw.directio.hive.parquet.DataModelWriteSupport
  • com.facebook.presto.hive.ParquetHiveRecordCursor$PrestoReadSupport
  • org.apache.drill.exec.store.parquet.ParquetRecordWriter
  • org.apache.drill.exec.store.parquet.ParquetScanBatchCreator
  • org.apache.drill.exec.store.parquet.TestFileGenerator
  • org.apache.drill.exec.store.parquet2.DrillParquetReader
  • org.apache.drill.exec.store.ParquetRecordReaderTest
  • org.apache.drill.ParquetSchemaMerge
  • org.apache.hadoop.hive.ql.io.parquet.convert.HiveSchemaConverter
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.