Package com.datasalt.pangool.io.Schema

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

113
114
115
116
117
118
119
120
121
122
123
124
   * @throws IOException
   */
  private void write(Schema destinationSchema, ITuple tuple,
      int[] translationTable, DataOutput output) throws IOException {
    for (int i = 0; i < destinationSchema.getFields().size(); i++) {
      Field field = destinationSchema.getField(i);
      Type fieldType = field.getType();
      Object element = tuple.get(translationTable[i]);
      try {
        switch(fieldType){
        case INT:
          WritableUtils.writeVInt(output, (Integer) element); break;
View Full Code Here
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
  }

  public void readFields(ITuple tuple, DataInput input) throws IOException {
    Schema schema = tuple.getSchema();
    for(int index = 0; index < schema.getFields().size(); 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 OBJECT: readCustomObject(input,tuple,field.getObjectClass(),index); break;
      default:
        throw new IOException("Not supported type:" + field.getType());
      }
    }
  }
View Full Code Here
126
127
128
129
130
131
132
133
134
135
      }
    }
  }

  private static Schema getPangoolTweetSchema() {
    Field tweetIdField = Field.create("tweet_id",Schema.Field.Type.INT);
    Field tweetHashTags = Fields.createAvroField("tweet_hashtags",
        getAvroStringArraySchema(),false);
    return new Schema("tweet",Arrays.asList(tweetIdField,tweetHashTags));
  }
View Full Code Here
133
134
135
136
137
138
139
140
141
        getAvroStringArraySchema(),false);
    return new Schema("tweet",Arrays.asList(tweetIdField,tweetHashTags));
  }
 
  private static Schema getPangoolRetweetSchema(){
    Field userId = Field.create("username",Schema.Field.Type.STRING);
    Field tweetId = Field.create("tweet_id",Schema.Field.Type.INT);
    return new Schema("retweet",Arrays.asList(userId,tweetId));
  }
View Full Code Here
124
125
126
127
128
129
130
131
  public AvroTopicalWordCount() {
    super("Usage: AvroTopicalWordCount [input_path] [output_path]");
  }

  static Schema getSchema() {
    Field avroField = Fields.createAvroField("my_avro",getAvroSchema(),false);
    return new Schema("schema",Arrays.asList(avroField));
  }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
    Schema intermediateSchema = mrConfig.getIntermediateSchemas().get(0);
    Criteria commonSortCriteria = mrConfig.getCommonCriteria();
    List<Field> commonFields = new ArrayList<Field>();
    for(SortElement sortElement : commonSortCriteria.getElements()) {
      String fieldName = sortElement.getName();
      Field field = checkFieldInAllSchemas(fieldName);
      commonFields.add(Field.cloneField(field, fieldName));
    }

    // adding the rest
    for(Field field : intermediateSchema.getFields()) {
      Map<String,String> aliases = mrConfig.getFieldAliases(intermediateSchema.getName());
      if(!containsField(field.getName(),commonFields,aliases))  {
          commonFields.add(field);
      }
    }
    this.commonSchema = new Schema("common", commonFields);
  }
View Full Code Here
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
  private void calculateMultipleSourcesSubSchemas() throws TupleMRException {
    Criteria commonSortCriteria = mrConfig.getCommonCriteria();
    List<Field> commonFields = new ArrayList<Field>();
    for(SortElement sortElement : commonSortCriteria.getElements()) {
      String fieldName = sortElement.getName();
      Field field = checkFieldInAllSchemas(fieldName);
     
      commonFields.add(Field.cloneField(field,fieldName));
    }

    this.commonSchema = new Schema("common", commonFields);
    this.specificSchemas = new ArrayList<Schema>();
    List<List<Field>> specificFieldsBySource = new ArrayList<List<Field>>();

    for(int schemaId = 0; schemaId < mrConfig.getNumIntermediateSchemas(); schemaId++) {
      Criteria specificCriteria = mrConfig.getSpecificOrderBys().get(schemaId);
      List<Field> specificFields = new ArrayList<Field>();
      if(specificCriteria != null) {
        for(SortElement sortElement : specificCriteria.getElements()) {
          String fieldName = sortElement.getName();
          Field field = checkFieldInSchema(fieldName, schemaId);
          specificFields.add(Field.cloneField(field,fieldName));
        }
      }
      specificFieldsBySource.add(specificFields);
    }

    for(int i = 0; i < mrConfig.getNumIntermediateSchemas(); i++) {
      Schema sourceSchema = mrConfig.getIntermediateSchema(i);
      List<Field> specificFields = specificFieldsBySource.get(i);
      for(Field field : sourceSchema.getFields()) {
        Map<String,String> sourceAliases = mrConfig.getFieldAliases(sourceSchema.getName());
        if(!containsField(field.getName(),commonSchema.getFields(),sourceAliases)
            && !containsField(field.getName(), specificFields,sourceAliases)) {
          specificFields.add(field);
        }
      }
      this.specificSchemas.add(new Schema("specific", specificFields));
    }
View Full Code Here
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
    }
    return false;
  }

  private Field checkFieldInAllSchemas(String name) throws TupleMRException {
    Field field = null;
    for(int i = 0; i < mrConfig.getIntermediateSchemas().size(); i++) {
      Field fieldInSource = checkFieldInSchema(name, i);
      if(field == null) {
        field = fieldInSource;
      } else if(field.getType() != fieldInSource.getType() || field.getObjectClass() != fieldInSource.getObjectClass()) {
        throw new TupleMRException("The type for field '" + name
            + "' is not the same in all the sources");
      }
    }
    return field;
View Full Code Here
240
241
242
243
244
245
246
247
248
249
250
  }

  private Field checkFieldInSchema(String fieldName, int schemaId)
      throws TupleMRException {
    Schema schema = mrConfig.getIntermediateSchema(schemaId);
    Field field = getFieldUsingAliases(schema,fieldName);
    if(field == null) {
      throw new TupleMRException("Field '" + fieldName + "' not present in source '"
          + schema.getName() + "' " + schema);
    }
    return field;
View Full Code Here
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
169
170
171
172
  public static void toRecord(ITuple tuple, Record record,
      DataOutputBuffer tmpOutputBuffer, HadoopSerialization ser) throws IOException {
    Schema pangoolSchema = tuple.getSchema();
    for(int i = 0; i < pangoolSchema.getFields().size(); i++) {
      Object obj = tuple.get(i);
      Field field = pangoolSchema.getField(i);
      switch(field.getType()){
      case INT:
      case LONG:
      case FLOAT:
      case BOOLEAN:
      case DOUBLE:
        record.put(i, obj); //optimistic
        break;
      case OBJECT:
        tmpOutputBuffer.reset();
        ser.ser(obj, tmpOutputBuffer);
        ByteBuffer buffer = ByteBuffer.wrap(tmpOutputBuffer.getData(), 0, tmpOutputBuffer.getLength());
        record.put(i, buffer);
        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());
      }
    }
  }
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.