Examples with MultipleFieldComparator org.apache.solr.client.solrj.io.comp.MultipleFieldComparator used on opensource projects

Example 1 with MultipleFieldComparator

use of org.apache.solr.client.solrj.io.comp.MultipleFieldComparator in project lucene-solr by apache.

the class SolrTable method handleSelectDistinctMapReduce.

private TupleStream handleSelectDistinctMapReduce(final String zkHost, final String collection, final Properties properties, final List<Map.Entry<String, Class>> fields, final String query, final List<Pair<String, String>> orders, final Bucket[] buckets, final String limit) throws IOException {
    int numWorkers = Integer.parseInt(properties.getProperty("numWorkers", "1"));
    String fl = getFields(fields);
    String sort = null;
    StreamEqualitor ecomp = null;
    StreamComparator comp = null;
    if (orders != null && orders.size() > 0) {
        StreamComparator[] adjustedSorts = adjustSorts(orders, buckets);
        // Because of the way adjustSorts works we know that each FieldComparator has a single
        // field name. For this reason we can just look at the leftFieldName
        FieldEqualitor[] fieldEqualitors = new FieldEqualitor[adjustedSorts.length];
        StringBuilder buf = new StringBuilder();
        for (int i = 0; i < adjustedSorts.length; i++) {
            FieldComparator fieldComparator = (FieldComparator) adjustedSorts[i];
            fieldEqualitors[i] = new FieldEqualitor(fieldComparator.getLeftFieldName());
            if (i > 0) {
                buf.append(",");
            }
            buf.append(fieldComparator.getLeftFieldName()).append(" ").append(fieldComparator.getOrder().toString());
        }
        sort = buf.toString();
        if (adjustedSorts.length == 1) {
            ecomp = fieldEqualitors[0];
            comp = adjustedSorts[0];
        } else {
            ecomp = new MultipleFieldEqualitor(fieldEqualitors);
            comp = new MultipleFieldComparator(adjustedSorts);
        }
    } else {
        StringBuilder sortBuf = new StringBuilder();
        FieldEqualitor[] equalitors = new FieldEqualitor[buckets.length];
        StreamComparator[] streamComparators = new StreamComparator[buckets.length];
        for (int i = 0; i < buckets.length; i++) {
            equalitors[i] = new FieldEqualitor(buckets[i].toString());
            streamComparators[i] = new FieldComparator(buckets[i].toString(), ComparatorOrder.ASCENDING);
            if (i > 0) {
                sortBuf.append(',');
            }
            sortBuf.append(buckets[i].toString()).append(" asc");
        }
        sort = sortBuf.toString();
        if (equalitors.length == 1) {
            ecomp = equalitors[0];
            comp = streamComparators[0];
        } else {
            ecomp = new MultipleFieldEqualitor(equalitors);
            comp = new MultipleFieldComparator(streamComparators);
        }
    }
    ModifiableSolrParams params = new ModifiableSolrParams();
    params.set(CommonParams.FL, fl);
    params.set(CommonParams.Q, query);
    params.set(CommonParams.WT, CommonParams.JAVABIN);
    //Always use the /export handler for Distinct Queries because it requires exporting full result sets.
    params.set(CommonParams.QT, "/export");
    if (numWorkers > 1) {
        params.set("partitionKeys", getPartitionKeys(buckets));
    }
    params.set(SORT, sort);
    TupleStream tupleStream = null;
    CloudSolrStream cstream = new CloudSolrStream(zkHost, collection, params);
    tupleStream = new UniqueStream(cstream, ecomp);
    if (numWorkers > 1) {
        // Do the unique in parallel
        // Maintain the sort of the Tuples coming from the workers.
        ParallelStream parallelStream = new ParallelStream(zkHost, collection, tupleStream, numWorkers, comp);
        StreamFactory factory = new StreamFactory().withFunctionName("search", CloudSolrStream.class).withFunctionName("parallel", ParallelStream.class).withFunctionName("unique", UniqueStream.class);
        parallelStream.setStreamFactory(factory);
        tupleStream = parallelStream;
    }
    if (limit != null) {
        tupleStream = new LimitStream(tupleStream, Integer.parseInt(limit));
    }
    return tupleStream;
}

Example 2 with MultipleFieldComparator

use of org.apache.solr.client.solrj.io.comp.MultipleFieldComparator in project lucene-solr by apache.

the class BiJoinStream method createIterationComparator.

private StreamComparator createIterationComparator(StreamEqualitor eq, StreamComparator comp) throws IOException {
    if (eq instanceof MultipleFieldEqualitor && comp instanceof MultipleFieldComparator) {
        // we know the comp is at least as long as the eq because we've already validated the tuple order
        StreamComparator[] compoundComps = new StreamComparator[((MultipleFieldEqualitor) eq).getEqs().length];
        for (int idx = 0; idx < compoundComps.length; ++idx) {
            StreamEqualitor sourceEqualitor = ((MultipleFieldEqualitor) eq).getEqs()[idx];
            StreamComparator sourceComparator = ((MultipleFieldComparator) comp).getComps()[idx];
            if (sourceEqualitor instanceof FieldEqualitor && sourceComparator instanceof FieldComparator) {
                FieldEqualitor fieldEqualitor = (FieldEqualitor) sourceEqualitor;
                FieldComparator fieldComparator = (FieldComparator) sourceComparator;
                compoundComps[idx] = new FieldComparator(fieldEqualitor.getLeftFieldName(), fieldEqualitor.getRightFieldName(), fieldComparator.getOrder());
            } else {
                throw new IOException("Failed to create an iteration comparator");
            }
        }
        return new MultipleFieldComparator(compoundComps);
    } else if (comp instanceof MultipleFieldComparator) {
        StreamEqualitor sourceEqualitor = eq;
        StreamComparator sourceComparator = ((MultipleFieldComparator) comp).getComps()[0];
        if (sourceEqualitor instanceof FieldEqualitor && sourceComparator instanceof FieldComparator) {
            FieldEqualitor fieldEqualitor = (FieldEqualitor) sourceEqualitor;
            FieldComparator fieldComparator = (FieldComparator) sourceComparator;
            return new FieldComparator(fieldEqualitor.getLeftFieldName(), fieldEqualitor.getRightFieldName(), fieldComparator.getOrder());
        } else {
            throw new IOException("Failed to create an iteration comparator");
        }
    } else {
        StreamEqualitor sourceEqualitor = eq;
        StreamComparator sourceComparator = comp;
        if (sourceEqualitor instanceof FieldEqualitor && sourceComparator instanceof FieldComparator) {
            FieldEqualitor fieldEqualitor = (FieldEqualitor) sourceEqualitor;
            FieldComparator fieldComparator = (FieldComparator) sourceComparator;
            return new FieldComparator(fieldEqualitor.getLeftFieldName(), fieldEqualitor.getRightFieldName(), fieldComparator.getOrder());
        } else {
            throw new IOException("Failed to create an iteration comparator");
        }
    }
}

Example 3 with MultipleFieldComparator

use of org.apache.solr.client.solrj.io.comp.MultipleFieldComparator in project lucene-solr by apache.

the class CloudSolrStream method parseComp.

private StreamComparator parseComp(String sort, String fl) throws IOException {
    String[] fls = fl.split(",");
    HashSet fieldSet = new HashSet();
    for (String f : fls) {
        //Handle spaces in the field list.
        fieldSet.add(f.trim());
    }
    String[] sorts = sort.split(",");
    StreamComparator[] comps = new StreamComparator[sorts.length];
    for (int i = 0; i < sorts.length; i++) {
        String s = sorts[i];
        //This should take into account spaces in the sort spec.
        String[] spec = s.trim().split("\\s+");
        if (spec.length != 2) {
            throw new IOException("Invalid sort spec:" + s);
        }
        String fieldName = spec[0].trim();
        String order = spec[1].trim();
        if (!fieldSet.contains(spec[0])) {
            throw new IOException("Fields in the sort spec must be included in the field list:" + spec[0]);
        }
        // if there's an alias for the field then use the alias
        if (null != fieldMappings && fieldMappings.containsKey(fieldName)) {
            fieldName = fieldMappings.get(fieldName);
        }
        comps[i] = new FieldComparator(fieldName, order.equalsIgnoreCase("asc") ? ComparatorOrder.ASCENDING : ComparatorOrder.DESCENDING);
    }
    if (comps.length > 1) {
        return new MultipleFieldComparator(comps);
    } else {
        return comps[0];
    }
}

Example 4 with MultipleFieldComparator

use of org.apache.solr.client.solrj.io.comp.MultipleFieldComparator in project lucene-solr by apache.

the class StreamFactory method constructComparator.

public StreamComparator constructComparator(String comparatorString, Class comparatorType) throws IOException {
    if (comparatorString.contains(",")) {
        String[] parts = comparatorString.split(",");
        StreamComparator[] comps = new StreamComparator[parts.length];
        for (int idx = 0; idx < parts.length; ++idx) {
            comps[idx] = constructComparator(parts[idx].trim(), comparatorType);
        }
        return new MultipleFieldComparator(comps);
    } else if (comparatorString.contains("=")) {
        // expected format is "left=right order"
        String[] parts = comparatorString.split("[ =]");
        if (parts.length < 3) {
            throw new IOException(String.format(Locale.ROOT, "Invalid comparator expression %s - expecting 'left=right order'", comparatorString));
        }
        String leftFieldName = null;
        String rightFieldName = null;
        String order = null;
        for (String part : parts) {
            // skip empty
            if (null == part || 0 == part.trim().length()) {
                continue;
            }
            // assign each in order
            if (null == leftFieldName) {
                leftFieldName = part.trim();
            } else if (null == rightFieldName) {
                rightFieldName = part.trim();
            } else {
                order = part.trim();
                // we're done, stop looping
                break;
            }
        }
        if (null == leftFieldName || null == rightFieldName || null == order) {
            throw new IOException(String.format(Locale.ROOT, "Invalid comparator expression %s - expecting 'left=right order'", comparatorString));
        }
        return (StreamComparator) createInstance(comparatorType, new Class[] { String.class, String.class, ComparatorOrder.class }, new Object[] { leftFieldName, rightFieldName, ComparatorOrder.fromString(order) });
    } else {
        // expected format is "field order"
        String[] parts = comparatorString.split(" ");
        if (2 != parts.length) {
            throw new IOException(String.format(Locale.ROOT, "Invalid comparator expression %s - expecting 'field order'", comparatorString));
        }
        String fieldName = parts[0].trim();
        String order = parts[1].trim();
        return (StreamComparator) createInstance(comparatorType, new Class[] { String.class, ComparatorOrder.class }, new Object[] { fieldName, ComparatorOrder.fromString(order) });
    }
}

Example 5 with MultipleFieldComparator

use of org.apache.solr.client.solrj.io.comp.MultipleFieldComparator in project lucene-solr by apache.

the class StreamingTest method testMergeStream.

@Test
public void testMergeStream() throws Exception {
    new UpdateRequest().add(id, "0", "a_s", "hello0", "a_i", "0", "a_f", "0").add(id, "2", "a_s", "hello2", "a_i", "2", "a_f", "0").add(id, "3", "a_s", "hello3", "a_i", "3", "a_f", "3").add(id, "4", "a_s", "hello4", "a_i", "4", "a_f", "4").add(id, "1", "a_s", "hello1", "a_i", "1", "a_f", "1").commit(cluster.getSolrClient(), COLLECTIONORALIAS);
    StreamContext streamContext = new StreamContext();
    SolrClientCache solrClientCache = new SolrClientCache();
    streamContext.setSolrClientCache(solrClientCache);
    try {
        //Test ascending
        SolrParams sParamsA = mapParams("q", "id:(4 1)", "fl", "id,a_s,a_i", "sort", "a_i asc");
        CloudSolrStream streamA = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsA);
        SolrParams sParamsB = mapParams("q", "id:(0 2 3)", "fl", "id,a_s,a_i", "sort", "a_i asc");
        CloudSolrStream streamB = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsB);
        MergeStream mstream = new MergeStream(streamA, streamB, new FieldComparator("a_i", ComparatorOrder.ASCENDING));
        mstream.setStreamContext(streamContext);
        List<Tuple> tuples = getTuples(mstream);
        assertEquals(5, tuples.size());
        assertOrder(tuples, 0, 1, 2, 3, 4);
        //Test descending
        sParamsA = mapParams("q", "id:(4 1)", "fl", "id,a_s,a_i", "sort", "a_i desc");
        streamA = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsA);
        sParamsB = mapParams("q", "id:(0 2 3)", "fl", "id,a_s,a_i", "sort", "a_i desc");
        streamB = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsB);
        mstream = new MergeStream(streamA, streamB, new FieldComparator("a_i", ComparatorOrder.DESCENDING));
        mstream.setStreamContext(streamContext);
        tuples = getTuples(mstream);
        assertEquals(5, tuples.size());
        assertOrder(tuples, 4, 3, 2, 1, 0);
        //Test compound sort
        sParamsA = mapParams("q", "id:(2 4 1)", "fl", "id,a_s,a_i,a_f", "sort", "a_f asc,a_i asc");
        streamA = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsA);
        sParamsB = mapParams("q", "id:(0 3)", "fl", "id,a_s,a_i,a_f", "sort", "a_f asc,a_i asc");
        streamB = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsB);
        mstream = new MergeStream(streamA, streamB, new MultipleFieldComparator(new FieldComparator("a_f", ComparatorOrder.ASCENDING), new FieldComparator("a_i", ComparatorOrder.ASCENDING)));
        mstream.setStreamContext(streamContext);
        tuples = getTuples(mstream);
        assertEquals(5, tuples.size());
        assertOrder(tuples, 0, 2, 1, 3, 4);
        sParamsA = mapParams("q", "id:(2 4 1)", "fl", "id,a_s,a_i,a_f", "sort", "a_f asc,a_i desc");
        streamA = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsA);
        sParamsB = mapParams("q", "id:(0 3)", "fl", "id,a_s,a_i,a_f", "sort", "a_f asc,a_i desc");
        streamB = new CloudSolrStream(zkHost, COLLECTIONORALIAS, sParamsB);
        mstream = new MergeStream(streamA, streamB, new MultipleFieldComparator(new FieldComparator("a_f", ComparatorOrder.ASCENDING), new FieldComparator("a_i", ComparatorOrder.DESCENDING)));
        mstream.setStreamContext(streamContext);
        tuples = getTuples(mstream);
        assertEquals(5, tuples.size());
        assertOrder(tuples, 2, 0, 1, 3, 4);
    } finally {
        solrClientCache.close();
    }
}