Package org.apache.crunch.types

Examples of org.apache.crunch.types.PTypeFamily

64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
   *          The right-side table of the join, whose contents will be fully
   *          read into memory
   * @return A table keyed on the join key, containing pairs of joined values
   */
  public static <K, U, V> PTable<K, Pair<U, V>> join(PTable<K, U> left, PTable<K, V> right) {
    PTypeFamily tf = left.getTypeFamily();
    Iterable<Pair<K, V>> iterable = right.materialize();

    if (iterable instanceof MaterializableIterable) {
      MaterializableIterable<Pair<K, V>> mi = (MaterializableIterable<Pair<K, V>>) iterable;
      MapsideJoinDoFn<K, U, V> mapJoinDoFn = new MapsideJoinDoFn<K, U, V>(mi.getPath().toString(),
          right.getPType());
      ParallelDoOptions.Builder optionsBuilder = ParallelDoOptions.builder();
      if (mi.isSourceTarget()) {
        optionsBuilder.sourceTargets((SourceTarget) mi.getSource());
      }
      return left.parallelDo("mapjoin", mapJoinDoFn,
          tf.tableOf(left.getKeyType(), tf.pairs(left.getValueType(), right.getValueType())),
          optionsBuilder.build());
    } else { // in-memory pipeline
      return left.parallelDo(new InMemoryJoinFn<K, U, V>(iterable),
          tf.tableOf(left.getKeyType(), tf.pairs(left.getValueType(), right.getValueType())));
    }
  }
View Full Code Here
114
115
116
117
118
119
120
121
122
123
124
125
   */
  public static <T> PCollection<T> reservorSample(
      PCollection<T> input,
      int sampleSize,
      Long seed) {
    PTypeFamily ptf = input.getTypeFamily();
    PType<Pair<T, Integer>> ptype = ptf.pairs(input.getPType(), ptf.ints());
    return weightedReservoirSample(
        input.parallelDo(new MapFn<T, Pair<T, Integer>>() {
          public Pair<T, Integer> map(T t) { return Pair.of(t, 1); }
        }, ptype),
        sampleSize,
View Full Code Here
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
   */
  public static <T, N extends Number> PCollection<T> weightedReservoirSample(
      PCollection<Pair<T, N>> input,
      int sampleSize,
      Long seed) {
    PTypeFamily ptf = input.getTypeFamily();
    PTable<Integer, Pair<T, N>> groupedIn = input.parallelDo(
        new MapFn<Pair<T, N>, Pair<Integer, Pair<T, N>>>() {
          @Override
          public Pair<Integer, Pair<T, N>> map(Pair<T, N> p) {
            return Pair.of(0, p);
          }
        }, ptf.tableOf(ptf.ints(), input.getPType()));
    int[] ss = new int[] { sampleSize };
    return groupedWeightedReservoirSample(groupedIn, ss, seed)
        .parallelDo(new MapFn<Pair<Integer, T>, T>() {
          @Override
          public T map(Pair<Integer, T> p) {
View Full Code Here
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
   */
  public static <T, N extends Number> PCollection<Pair<Integer, T>> groupedWeightedReservoirSample(
      PTable<Integer, Pair<T, N>> input,
      int[] sampleSizes,
      Long seed) {
    PTypeFamily ptf = input.getTypeFamily();
    PType<T> ttype = (PType<T>) input.getPTableType().getValueType().getSubTypes().get(0);
    PTableType<Integer, Pair<Double, T>> ptt = ptf.tableOf(ptf.ints(),
        ptf.pairs(ptf.doubles(), ttype));
   
    return input.parallelDo(new ReservoirSampleFn<T, N>(sampleSizes, seed, ttype), ptt)
        .groupByKey(1)
        .combineValues(new WRSCombineFn<T>(sampleSizes, ttype))
        .parallelDo(new MapFn<Pair<Integer, Pair<Double, T>>, Pair<Integer, T>>() {
          @Override
          public Pair<Integer, T> map(Pair<Integer, Pair<Double, T>> p) {
            return Pair.of(p.first(), p.second().second());
          }
        }, ptf.pairs(ptf.ints(), ttype));
  }
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
   * @return A new {@code PCollection} that contains the unique elements of the input
   */
  public static <S> PCollection<S> distinct(PCollection<S> input, int flushEvery) {
    Preconditions.checkArgument(flushEvery > 0);
    PType<S> pt = input.getPType();
    PTypeFamily ptf = pt.getFamily();
    return input
        .parallelDo("pre-distinct", new PreDistinctFn<S>(flushEvery, pt), ptf.tableOf(pt, ptf.nulls()))
        .groupByKey()
        .parallelDo("post-distinct", new PostDistinctFn<S>(), pt);
  }
View Full Code Here
50
51
52
53
54
55
56
57
58
59
60
61
62
  /**
   * Returns a {@code PTable} that contains the unique elements of this collection mapped to a count
   * of their occurrences.
   */
  public static <S> PTable<S, Long> count(PCollection<S> collect) {
    PTypeFamily tf = collect.getTypeFamily();
    return collect.parallelDo("Aggregate.count", new MapFn<S, Pair<S, Long>>() {
      public Pair<S, Long> map(S input) {
        return Pair.of(input, 1L);
      }
    }, tf.tableOf(collect.getPType(), tf.longs())).groupByKey()
        .combineValues(Aggregators.SUM_LONGS());
  }
View Full Code Here
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
   * @param collect The PCollection whose elements should be counted.
   * @param <S> The type of the PCollection.
   * @return A {@code PObject} containing the number of elements in the {@code PCollection}.
   */
  public static <S> PObject<Long> length(PCollection<S> collect) {
    PTypeFamily tf = collect.getTypeFamily();
    PTable<Integer, Long> countTable = collect
        .parallelDo("Aggregate.count", new MapFn<S, Pair<Integer, Long>>() {
          public Pair<Integer, Long> map(S input) {
            return Pair.of(1, 1L);
          }
        }, tf.tableOf(tf.ints(), tf.longs()))
        .groupByKey(GroupingOptions.builder().numReducers(1).build())
        .combineValues(Aggregators.SUM_LONGS());
    PCollection<Long> count = countTable.values();
    return new FirstElementPObject<Long>(count);
  }
View Full Code Here
154
155
156
157
158
159
160
161
162
163
164
165
166
167
      }
    }
  }

  public static <K, V> PTable<K, V> top(PTable<K, V> ptable, int limit, boolean maximize) {
    PTypeFamily ptf = ptable.getTypeFamily();
    PTableType<K, V> base = ptable.getPTableType();
    PType<Pair<K, V>> pairType = ptf.pairs(base.getKeyType(), base.getValueType());
    PTableType<Integer, Pair<K, V>> inter = ptf.tableOf(ptf.ints(), pairType);
    return ptable.parallelDo("top" + limit + "map", new TopKFn<K, V>(limit, maximize), inter)
        .groupByKey(1).combineValues(new TopKCombineFn<K, V>(limit, maximize))
        .parallelDo("top" + limit + "reduce", new DoFn<Pair<Integer, Pair<K, V>>, Pair<K, V>>() {
          public void process(Pair<Integer, Pair<K, V>> input, Emitter<Pair<K, V>> emitter) {
            emitter.emit(input.second());
View Full Code Here
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
    Class<S> clazz = collect.getPType().getTypeClass();
    if (!clazz.isPrimitive() && !Comparable.class.isAssignableFrom(clazz)) {
      throw new IllegalArgumentException("Can only get max for Comparable elements, not for: "
          + collect.getPType().getTypeClass());
    }
    PTypeFamily tf = collect.getTypeFamily();
    PCollection<S> maxCollect = PTables.values(collect
        .parallelDo("max", new DoFn<S, Pair<Boolean, S>>() {
          private transient S max = null;

          public void process(S input, Emitter<Pair<Boolean, S>> emitter) {
            if (max == null || ((Comparable<S>) max).compareTo(input) < 0) {
              max = input;
            }
          }

          public void cleanup(Emitter<Pair<Boolean, S>> emitter) {
            if (max != null) {
              emitter.emit(Pair.of(true, max));
            }
          }
        }, tf.tableOf(tf.booleans(), collect.getPType())).groupByKey(1)
        .combineValues(new CombineFn<Boolean, S>() {
          public void process(Pair<Boolean, Iterable<S>> input, Emitter<Pair<Boolean, S>> emitter) {
            S max = null;
            for (S v : input.second()) {
              if (max == null || ((Comparable<S>) max).compareTo(v) < 0) {
View Full Code Here
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
    Class<S> clazz = collect.getPType().getTypeClass();
    if (!clazz.isPrimitive() && !Comparable.class.isAssignableFrom(clazz)) {
      throw new IllegalArgumentException("Can only get min for Comparable elements, not for: "
          + collect.getPType().getTypeClass());
    }
    PTypeFamily tf = collect.getTypeFamily();
    PCollection<S> minCollect = PTables.values(collect
        .parallelDo("min", new DoFn<S, Pair<Boolean, S>>() {
          private transient S min = null;

          public void process(S input, Emitter<Pair<Boolean, S>> emitter) {
            if (min == null || ((Comparable<S>) min).compareTo(input) > 0) {
              min = input;
            }
          }

          public void cleanup(Emitter<Pair<Boolean, S>> emitter) {
            if (min != null) {
              emitter.emit(Pair.of(false, min));
            }
          }
        }, tf.tableOf(tf.booleans(), collect.getPType())).groupByKey(1)
        .combineValues(new CombineFn<Boolean, S>() {
          public void process(Pair<Boolean, Iterable<S>> input, Emitter<Pair<Boolean, S>> emitter) {
            S min = null;
            for (S v : input.second()) {
              if (min == null || ((Comparable<S>) min).compareTo(v) > 0) {
View Full Code Here
TOP

Related Classes of org.apache.crunch.types.PTypeFamily

  • org.apache.crunch.contrib.bloomfilter.BloomFilterFactory
  • org.apache.crunch.io.At
  • org.apache.crunch.io.From
  • org.apache.crunch.lib.Aggregate
  • org.apache.crunch.lib.Cartesian
  • org.apache.crunch.lib.Cogroup
  • org.apache.crunch.lib.Distinct
  • org.apache.crunch.lib.Join
  • org.apache.crunch.lib.join.DefaultJoinStrategy
  • org.apache.crunch.lib.join.MapsideJoin
    • 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.