Examples with GroupingOptions org.apache.crunch.GroupingOptions used on opensource projects

Example 6 with GroupingOptions

use of org.apache.crunch.GroupingOptions in project crunch by cloudera.

the class Sort method sort.

/**
   * Sorts the {@link PCollection} using the natural ordering of its elements
   * in the order specified.
   * 
   * @return a {@link PCollection} representing the sorted collection.
   */
public static <T> PCollection<T> sort(PCollection<T> collection, Order order) {
    PTypeFamily tf = collection.getTypeFamily();
    PTableType<T, Void> type = tf.tableOf(collection.getPType(), tf.nulls());
    Configuration conf = collection.getPipeline().getConfiguration();
    GroupingOptions options = buildGroupingOptions(conf, tf, collection.getPType(), order);
    PTable<T, Void> pt = collection.parallelDo("sort-pre", new DoFn<T, Pair<T, Void>>() {

        @Override
        public void process(T input, Emitter<Pair<T, Void>> emitter) {
            emitter.emit(Pair.of(input, (Void) null));
        }
    }, type);
    PTable<T, Void> sortedPt = pt.groupByKey(options).ungroup();
    return sortedPt.parallelDo("sort-post", new DoFn<Pair<T, Void>, T>() {

        @Override
        public void process(Pair<T, Void> input, Emitter<T> emitter) {
            emitter.emit(input.first());
        }
    }, collection.getPType());
}