@Override
public int hashCode() {
int result = dimension.hashCode();
result = 31 * result + (value != null ? value.hashCode() : 0);
result = 31 * result + (extractionFn != null ? extractionFn.hashCode() : 0);
return result;
}
@Test
public void testExtraction() {
String regex = ".*[Tt][Oo].*";
ExtractionFn extractionFn = new MatchingDimExtractionFn(regex);
List<String> expected = Arrays.asList("Quito", "Tokyo", "Stockholm", "Pretoria", "Wellington");
Set<String> extracted = Sets.newHashSet();
for (String str : testStrings) {
String res = extractionFn.apply(str);
if (res != null) {
extracted.add(res);
}
}
Assert.assertEquals(5, extracted.size());
for (String str : extracted) {
Assert.assertTrue(expected.contains(str));
}
}
@Override
public byte[] getCacheKey() {
byte[] dimensionBytes = StringUtils.toUtf8(dimension);
byte[] valueBytes = (value == null) ? new byte[] {} : StringUtils.toUtf8(value);
byte[] extractionFnBytes = extractionFn == null ? new byte[0] : extractionFn.getCacheKey();
return ByteBuffer.allocate(
3 + dimensionBytes.length + valueBytes.length + extractionFnBytes.length)
.put(DimFilterCacheHelper.SELECTOR_CACHE_ID)
.put(dimensionBytes)
.put(DimFilterCacheHelper.STRING_SEPARATOR)
.put(valueBytes)
.put(DimFilterCacheHelper.STRING_SEPARATOR)
.put(extractionFnBytes)
.array();
}
@Override
public boolean equals(Object o) {
if (this == o) {
return true;
}
if (o == null || getClass() != o.getClass()) {
return false;
}
SelectorDimFilter that = (SelectorDimFilter) o;
if (!dimension.equals(that.dimension)) {
return false;
}
if (value != null ? !value.equals(that.value) : that.value != null) {
return false;
}
return extractionFn != null
? extractionFn.equals(that.extractionFn)
: that.extractionFn == null;
}
@Override
public Sequence<Result<SearchResultValue>> run(
final Query<Result<SearchResultValue>> input, Map<String, Object> responseContext) {
if (!(input instanceof SearchQuery)) {
throw new ISE("Got a [%s] which isn't a %s", input.getClass(), SearchQuery.class);
}
final SearchQuery query = (SearchQuery) input;
final Filter filter = Filters.convertDimensionFilters(query.getDimensionsFilter());
final List<DimensionSpec> dimensions = query.getDimensions();
final SearchQuerySpec searchQuerySpec = query.getQuery();
final int limit = query.getLimit();
final boolean descending = query.isDescending();
// Closing this will cause segfaults in unit tests.
final QueryableIndex index = segment.asQueryableIndex();
if (index != null) {
final TreeMap<SearchHit, MutableInt> retVal =
Maps.newTreeMap(query.getSort().getComparator());
Iterable<DimensionSpec> dimsToSearch;
if (dimensions == null || dimensions.isEmpty()) {
dimsToSearch =
Iterables.transform(index.getAvailableDimensions(), Druids.DIMENSION_IDENTITY);
} else {
dimsToSearch = dimensions;
}
final BitmapFactory bitmapFactory = index.getBitmapFactoryForDimensions();
final ImmutableBitmap baseFilter =
filter == null
? null
: filter.getBitmapIndex(new ColumnSelectorBitmapIndexSelector(bitmapFactory, index));
for (DimensionSpec dimension : dimsToSearch) {
final Column column = index.getColumn(dimension.getDimension());
if (column == null) {
continue;
}
final BitmapIndex bitmapIndex = column.getBitmapIndex();
ExtractionFn extractionFn = dimension.getExtractionFn();
if (extractionFn == null) {
extractionFn = IdentityExtractionFn.getInstance();
}
if (bitmapIndex != null) {
for (int i = 0; i < bitmapIndex.getCardinality(); ++i) {
String dimVal = Strings.nullToEmpty(extractionFn.apply(bitmapIndex.getValue(i)));
if (!searchQuerySpec.accept(dimVal)) {
continue;
}
ImmutableBitmap bitmap = bitmapIndex.getBitmap(i);
if (baseFilter != null) {
bitmap = bitmapFactory.intersection(Arrays.asList(baseFilter, bitmap));
}
if (bitmap.size() > 0) {
MutableInt counter = new MutableInt(bitmap.size());
MutableInt prev =
retVal.put(new SearchHit(dimension.getOutputName(), dimVal), counter);
if (prev != null) {
counter.add(prev.intValue());
}
if (retVal.size() >= limit) {
return makeReturnResult(limit, retVal);
}
}
}
}
}
return makeReturnResult(limit, retVal);
}
final StorageAdapter adapter = segment.asStorageAdapter();
if (adapter == null) {
log.makeAlert("WTF!? Unable to process search query on segment.")
.addData("segment", segment.getIdentifier())
.addData("query", query)
.emit();
throw new ISE(
"Null storage adapter found. Probably trying to issue a query against a segment being memory unmapped.");
}
final Iterable<DimensionSpec> dimsToSearch;
if (dimensions == null || dimensions.isEmpty()) {
dimsToSearch =
Iterables.transform(adapter.getAvailableDimensions(), Druids.DIMENSION_IDENTITY);
} else {
dimsToSearch = dimensions;
}
final Sequence<Cursor> cursors =
adapter.makeCursors(filter, segment.getDataInterval(), QueryGranularity.ALL, descending);
final TreeMap<SearchHit, MutableInt> retVal =
cursors.accumulate(
Maps.<SearchHit, SearchHit, MutableInt>newTreeMap(query.getSort().getComparator()),
new Accumulator<TreeMap<SearchHit, MutableInt>, Cursor>() {
@Override
public TreeMap<SearchHit, MutableInt> accumulate(
TreeMap<SearchHit, MutableInt> set, Cursor cursor) {
if (set.size() >= limit) {
return set;
}
Map<String, DimensionSelector> dimSelectors = Maps.newHashMap();
for (DimensionSpec dim : dimsToSearch) {
dimSelectors.put(dim.getOutputName(), cursor.makeDimensionSelector(dim));
}
while (!cursor.isDone()) {
for (Map.Entry<String, DimensionSelector> entry : dimSelectors.entrySet()) {
final DimensionSelector selector = entry.getValue();
if (selector != null) {
final IndexedInts vals = selector.getRow();
for (int i = 0; i < vals.size(); ++i) {
final String dimVal = selector.lookupName(vals.get(i));
if (searchQuerySpec.accept(dimVal)) {
MutableInt counter = new MutableInt(1);
MutableInt prev = set.put(new SearchHit(entry.getKey(), dimVal), counter);
if (prev != null) {
counter.add(prev.intValue());
}
if (set.size() >= limit) {
return set;
}
}
}
}
}
cursor.advance();
}
return set;
}
});
return makeReturnResult(limit, retVal);
}