Package com.datasalt.pangool.io.Schema

Examples of com.datasalt.pangool.io.Schema.Field

114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
  }

  public int compare(Schema schema, Criteria c, ITuple w1, int[] index1, ITuple w2,
                     int[] index2, Serializer[] serializers) {
    for (int i = 0; i < c.getElements().size(); i++) {
      Field field = schema.getField(i);
      SortElement e = c.getElements().get(i);
      Object o1 = w1.get(index1[i]);
      Object o2 = w2.get(index2[i]);

      // Handling with null values
      if (o1 == null || o2 == null) {
        int cmp = nullCompare(o1, o2, e);
        if (cmp != 0) {
          return cmp;
        } else {
          continue;
        }
      }

      // At this point we know that both values are not null.
      Serializer serializer = (serializers == null) ? null : serializers[i];
      int comparison = compareObjects(o1, o2, e.getCustomComparator(), field.getType(), serializer);
      if (comparison != 0) {
        return (e.getOrder() == Order.ASC ? comparison : -comparison);
      }
    }
    return 0;
View Full Code Here
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
      o.offset1 += n.nulls1.deser(b1, s1);
      o.offset2 += n.nulls2.deser(b2, s2);
    }

    for (int depth = 0; depth < criteria.getElements().size(); depth++) {
      Field field = schema.getField(depth);
      Field.Type type = field.getType();
      SortElement sortElement = criteria.getElements().get(depth);
      Order sort = sortElement.getOrder();
      RawComparator comparator = sortElement.getCustomComparator();

      // Control for nulls, if field is nullable.
      if (field.isNullable()) {
        Criteria.NullOrder nullOrder = sortElement.getNullOrder();
        if (n.nulls1.isSet(schema.getNullablePositionFromIndex(depth))) {
          if (n.nulls2.isSet(schema.getNullablePositionFromIndex(depth))) {
            // Both are null, so both are equal. No space is used. Continue.
            continue;
View Full Code Here
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
      }
      nulls.ser(out);
    }

    for (int i = 0; i < destinationSchema.getFields().size(); i++) {
      Field field = destinationSchema.getField(i);
      Type fieldType = field.getType();
      Object element = valueAt(i, tuple, translationTable);
      if (element == null) {
        if (field.isNullable()) {
          // Nullable null fields don't need serialization.
          continue;
        } else {
          raiseUnexpectedNullException(field, element);
        }
View Full Code Here
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
    Schema pangoolSchema = tuple.getSchema();
    for(org.apache.avro.Schema.Field avroField : avroSchema.getFields()) {
      int pos = avroField.pos();
      Object objRecord = record.get(pos);
      Field pangoolField = pangoolSchema.getField(pos);
      switch(pangoolField.getType()) {
      case INT:
      case LONG:
      case BOOLEAN:
      case FLOAT:
      case DOUBLE:
        tuple.set(pos, objRecord); // very optimistic
        break;
      case STRING: {
        if(!(tuple.get(pos) instanceof Utf8)) {
          tuple.set(pos, new com.datasalt.pangool.io.Utf8());
        }
        com.datasalt.pangool.io.Utf8 utf8 = (com.datasalt.pangool.io.Utf8) tuple.get(pos);
        if(objRecord instanceof String) {
          utf8.set((String) objRecord);
        } else if(objRecord instanceof Utf8) {
          Utf8 avroUtf8 = (Utf8) objRecord;
          utf8.set(avroUtf8.getBytes(), 0, avroUtf8.getByteLength());
        } else {
          throw new IOException("Not supported avro field " + org.apache.avro.Schema.Type.STRING + " with instance "
              + objRecord.getClass().getName());
        }
        break;
      }
      case ENUM: {
        Class clazz = pangoolField.getObjectClass();
        Enum e = Enum.valueOf(clazz, objRecord.toString());
        tuple.set(pos, e);
        break;
      }
      case BYTES:
        tuple.set(pos, objRecord); // TODO FIXME this should copy bytes really, not reference!
        break;
      case OBJECT:
        Deserializer customDeser = customDeserializers[pos];
        if(objRecord instanceof byte[]) {
          inputBuffer.reset((byte[]) objRecord, ((byte[]) objRecord).length);
        } else if(objRecord instanceof ByteBuffer) {
          ByteBuffer buffer = (ByteBuffer) objRecord;
          int offset = buffer.arrayOffset() + buffer.position();
          int length = buffer.limit() - buffer.position();
          inputBuffer.reset(buffer.array(), offset, length);
        } else {
          throw new PangoolRuntimeException("Can't convert to OBJECT from instance " + objRecord.getClass());
        }
        if(customDeser != null) {
          customDeser.open(inputBuffer);
          tuple.set(pos, customDeser.deserialize(tuple.get(pos))); // TODO FIXME avro deserializer shouldn't reuse
                                                                   // objects sometimes (UNION ?)
          customDeser.close(); // TODO is this ok ?
        } else {
          // no custom deser , then use Hadoop serializers registered in "io.serializations"
          Class clazz = pangoolField.getObjectClass();
          if(tuple.get(pos) == null || tuple.get(pos).getClass() != clazz) {
            tuple.set(pos, ReflectionUtils.newInstance(clazz, conf));
          }
          hadoopSer.deser(tuple.get(pos), inputBuffer);
        }
View Full Code Here
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
          "contains schema not expected." +
          "Expected schema '"+ pangoolSchema + " and actual: " + tuple.getSchema());
    }
    for(int i = 0; i < pangoolSchema.getFields().size(); i++) {
      Object obj = tuple.get(i);
      Field field = pangoolSchema.getField(i);
      if (obj == null){
        throw new IOException("Field '"
      + field.getName() + "' can't be null in tuple:" + tuple);
      }
     
      switch(field.getType()){
      case INT:
      case LONG:
      case FLOAT:
      case BOOLEAN:
      case DOUBLE:
      case BYTES:
        record.put(i, obj); //optimistic
        break;
      case OBJECT:
        Serializer customSer = customSerializers[i];
        DataOutputBuffer buffer = buffers[i];
        buffer.reset();
        if (customSer != null){
          customSer.open(buffer);
          customSer.serialize(obj);
          customSer.close(); //TODO is this safe ?
        } else {
          hadoopSer.ser(obj, buffer);
        }
        //TODO this byteBuffer instances should be cached and reused
        ByteBuffer byteBuffer = ByteBuffer.wrap(buffer.getData(), 0,buffer.getLength());
        record.put(i, byteBuffer);
        break;
      case ENUM:
        record.put(i,obj.toString());
        break;
      case STRING:
        record.put(i,new Utf8(obj.toString())); //could be directly String ?
        break;
      default:
          throw
          new IOException("Not correspondence to Avro type from Pangool type " + field.getType());
      }
    }
    return record;
  }
View Full Code Here
96
97
98
99
100
101
102
103
104
105
106
   * or deserialized
   * @return
   */
  public static Field createAvroField(String name,
       org.apache.avro.Schema avroSchema,  boolean isReflect){
    Field field = Field.createObject(name,Object.class);
    field.setObjectSerialization(AvroFieldSerialization.class);
    field.addProp("avro.schema",avroSchema.toString());
    field.addProp("avro.reflection",Boolean.toString(isReflect));
    return field;
  }
View Full Code Here
111
112
113
114
115
116
117
118
119
120
   * @param name Field's name
   * @param schema The schema of the field
   * @return
   */
  public static Field createTupleField(String name, Schema schema) {
    Field field = Field.createObject(name,Object.class);
    field.setObjectSerialization(TupleFieldSerialization.class);
    field.addProp("schema", schema.toString());
    return field;
  }
View Full Code Here
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127

  public int compare(Schema schema, Criteria c, ITuple w1, int[] index1, ITuple w2,
      int[] index2,Serializer[] serializers) {
    for(int i = 0; i < c.getElements().size(); i++) {
      Field field = schema.getField(i);
      SortElement e = c.getElements().get(i);
      Object o1 = w1.get(index1[i]);
      Object o2 = w2.get(index2[i]);
      Serializer serializer = (serializers == null) ? null : serializers[i];
      int comparison = compareObjects(o1, o2, e.getCustomComparator(), field.getType(),serializer);
      if(comparison != 0) {
        return(e.getOrder() == Order.ASC ? comparison : -comparison);
      }
    }
    return 0;
View Full Code Here
254
255
256
257
258
259
260
261
262
263
264
265
  protected int compare(byte[] b1, int s1, byte[] b2, int s2, Schema schema,
      Criteria criteria, Offsets o) throws IOException {
    o.offset1 = s1;
    o.offset2 = s2;
    for(int depth = 0; depth < criteria.getElements().size(); depth++) {
      Field field = schema.getField(depth);
      Field.Type type = field.getType();
      SortElement sortElement = criteria.getElements().get(depth);
      Order sort = sortElement.getOrder();
      RawComparator comparator = sortElement.getCustomComparator();

      if(comparator != null) {
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
  public void readFields(ITuple tuple, Deserializer[] customDeserializers) throws IOException {
    Schema schema = tuple.getSchema();
    for(int index = 0; index < schema.getFields().size(); index++) {
      Deserializer customDeser = customDeserializers[index];
      Field field = schema.getField(index);
      switch(field.getType()) {
      case INT:
        tuple.set(index, WritableUtils.readVInt(input));
        break;
      case LONG:
        tuple.set(index, WritableUtils.readVLong(input));
        break;
      case DOUBLE:
        tuple.set(index, input.readDouble());
        break;
      case FLOAT:
        tuple.set(index, input.readFloat());
        break;
      case STRING:
        readUtf8(input, tuple, index);
        break;
      case BOOLEAN:
        byte b = input.readByte();
        tuple.set(index, (b != 0));
        break;
      case ENUM:
        readEnum(input, tuple, field.getObjectClass(), index);
        break;
      case BYTES:
        readBytes(input, tuple, index);
        break;
      case OBJECT:
        readCustomObject(input, tuple, field.getObjectClass(), index, customDeser);
        break;
      default:
        throw new IOException("Not supported type:" + field.getType());
      }
    }
  }
View Full Code Here
TOP

Related Classes of com.datasalt.pangool.io.Schema.Field

  • com.datasalt.pangool.BaseTest
  • com.datasalt.pangool.examples.avro.AvroCustomSerializationJob
  • com.datasalt.pangool.examples.avro.AvroTopicalWordCount
  • com.datasalt.pangool.examples.avro.AvroTweetsJoin
  • com.datasalt.pangool.io.Fields
  • com.datasalt.pangool.io.Tuple
  • com.datasalt.pangool.tuplemr.mapred.lib.input.TupleTextInputFormat$TupleTextInputReader
  • com.datasalt.pangool.tuplemr.mapred.SortComparator
  • com.datasalt.pangool.tuplemr.mapred.TestComparators
  • com.datasalt.pangool.tuplemr.serialization.SimpleTupleDeserializer
    • 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.