@Override
public void collect(int doc, long owningBucketOrdinal) throws IOException {
final int numOrds = segmentOrdinals.setDocument(doc);
for (int i = 0; i < numOrds; i++) {
final long segmentOrd = segmentOrdinals.nextOrd();
current.increment(segmentOrd, 1);
}
}
@Override
public void collect(int doc, long owningBucketOrdinal) throws IOException {
final int numOrds = globalOrdinals.setDocument(doc);
for (int i = 0; i < numOrds; i++) {
final long globalOrd = globalOrdinals.nextOrd();
collectExistingBucket(doc, globalOrd);
}
}
@Override
public int setDocument(int docId) {
int numDocs = inner.setDocument(docId);
buffer = ArrayUtil.grow(buffer, numDocs);
bufferSlot = 0;
int numAcceptedOrds = 0;
for (int slot = 0; slot < numDocs; slot++) {
long ord = inner.nextOrd();
if (accepted.get(ord)) {
buffer[numAcceptedOrds] = ord;
numAcceptedOrds++;
}
}
return numAcceptedOrds;
}
@Override
public void setNextReader(AtomicReaderContext reader) {
if (segmentOrdinals != null && segmentOrdinals.getMaxOrd() != globalOrdinals.getMaxOrd()) {
mapSegmentCountsToGlobalCounts();
}
globalValues = valuesSource.globalBytesValues();
globalOrdinals = globalValues.ordinals();
BytesValues.WithOrdinals bytesValues = valuesSource.bytesValues();
segmentOrdinals = bytesValues.ordinals();
if (segmentOrdinals.getMaxOrd() != globalOrdinals.getMaxOrd()) {
current = segmentDocCounts;
} else {
current = getDocCounts();
}
}
@Override
public long getOrd(int docId) {
long ord = inner.getOrd(docId);
if (accepted.get(ord)) {
return currentOrd = ord;
} else {
return currentOrd = Ordinals.MISSING_ORDINAL;
}
}
@Override
public void setNextReader(AtomicReaderContext context) throws IOException {
values = indexFieldData.load(context).getBytesValues(parentType);
if (values != null) {
ordinals = values.ordinals();
final long maxOrd = ordinals.getMaxOrd();
if (parentIdsIndex == null) {
parentIdsIndex = bigArrays.newLongArray(BigArrays.overSize(maxOrd), false);
} else if (parentIdsIndex.size() < maxOrd) {
parentIdsIndex = bigArrays.grow(parentIdsIndex, maxOrd);
}
parentIdsIndex.fill(0, maxOrd, -1L);
}
}
@Override
public void collect(int doc) throws IOException {
if (values != null) {
long ord = ordinals.getOrd(doc);
long parentIdx = parentIdsIndex.get(ord);
if (parentIdx < 0) {
final BytesRef bytes = values.getValueByOrd(ord);
final int hash = values.currentValueHash();
parentIdx = parentIds.add(bytes, hash);
if (parentIdx < 0) {
parentIdx = -parentIdx - 1;
doScore(parentIdx);
} else {
scores = bigArrays.grow(scores, parentIdx + 1);
scores.set(parentIdx, scorer.score());
}
parentIdsIndex.set(ord, parentIdx);
} else {
doScore(parentIdx);
}
}
}
@Override
public boolean isMultiValued() {
return inner.isMultiValued();
}
@Override
public long getMaxOrd() {
return inner.getMaxOrd();
}
@Override
protected void doPostCollection() {
if (segmentOrdinals.getMaxOrd() != globalOrdinals.getMaxOrd()) {
mapSegmentCountsToGlobalCounts();
}
}
@Override
public InternalAggregation buildAggregation(long owningBucketOrdinal) {
if (globalOrdinals == null) { // no context in this reader
return buildEmptyAggregation();
}
final int size;
if (bucketCountThresholds.getMinDocCount() == 0) {
// if minDocCount == 0 then we can end up with more buckets then maxBucketOrd() returns
size = (int) Math.min(globalOrdinals.getMaxOrd(), bucketCountThresholds.getShardSize());
} else {
size = (int) Math.min(maxBucketOrd(), bucketCountThresholds.getShardSize());
}
BucketPriorityQueue ordered = new BucketPriorityQueue(size, order.comparator(this));
OrdBucket spare = new OrdBucket(-1, 0, null);
for (long globalTermOrd = Ordinals.MIN_ORDINAL;
globalTermOrd < globalOrdinals.getMaxOrd();
++globalTermOrd) {
if (includeExclude != null && !acceptedGlobalOrdinals.get(globalTermOrd)) {
continue;
}
final long bucketOrd = getBucketOrd(globalTermOrd);
final long bucketDocCount = bucketOrd < 0 ? 0 : bucketDocCount(bucketOrd);
if (bucketCountThresholds.getMinDocCount() > 0 && bucketDocCount == 0) {
continue;
}
spare.globalOrd = globalTermOrd;
spare.bucketOrd = bucketOrd;
spare.docCount = bucketDocCount;
if (bucketCountThresholds.getShardMinDocCount() <= spare.docCount) {
spare = (OrdBucket) ordered.insertWithOverflow(spare);
if (spare == null) {
spare = new OrdBucket(-1, 0, null);
}
}
}
// Get the top buckets
final InternalTerms.Bucket[] list = new InternalTerms.Bucket[ordered.size()];
long survivingBucketOrds[] = new long[ordered.size()];
for (int i = ordered.size() - 1; i >= 0; --i) {
final OrdBucket bucket = (OrdBucket) ordered.pop();
survivingBucketOrds[i] = bucket.bucketOrd;
BytesRef scratch = new BytesRef();
copy(globalValues.getValueByOrd(bucket.globalOrd), scratch);
list[i] = new StringTerms.Bucket(scratch, bucket.docCount, null);
list[i].bucketOrd = bucket.bucketOrd;
}
// replay any deferred collections
runDeferredCollections(survivingBucketOrds);
// Now build the aggs
for (int i = 0; i < list.length; i++) {
Bucket bucket = list[i];
bucket.aggregations =
bucket.docCount == 0 ? bucketEmptyAggregations() : bucketAggregations(bucket.bucketOrd);
}
return new StringTerms(
name,
order,
bucketCountThresholds.getRequiredSize(),
bucketCountThresholds.getMinDocCount(),
Arrays.asList(list));
}