/**
* For use within an iterator stack. Apply an appropriate column filter based on the input string.
* Four modes of operation: 1. Null or blank ("") `colFilter`: do nothing. 2. No ranges
* `colFilter`: use Accumulo system ColumnQualifierFilter. 3. Singleton range `colFilter`: use
* Accumulo user ColumnSliceFilter. 4. Multi-range `colFilter`: use Graphulo D4mRangeFilter.
*
* @param colFilter column filter string
* @param skvi Parent / source iterator
* @return SKVI with appropriate filter iterators placed in front of it.
*/
public static SortedKeyValueIterator<Key, Value> applyGeneralColumnFilter(
String colFilter, SortedKeyValueIterator<Key, Value> skvi, IteratorEnvironment env)
throws IOException {
if (colFilter == null || colFilter.isEmpty()) return skvi;
int pos1 = colFilter.indexOf(':');
if (pos1 == -1) { // no ranges - collection of singleton texts
Set<Column> colset = new HashSet<>();
for (Text text : GraphuloUtil.d4mRowToTexts(colFilter)) {
byte[] by = text.copyBytes();
// log.debug("Printing characters of string TEXT LIM: "+ Key.toPrintableString(by, 0,
// text.getLength(), 100));
// log.debug("Printing characters of string TEXT : "+ Key.toPrintableString(by, 0,
// by.length, 100));
colset.add(new Column(EMPTY_BYTES, text.copyBytes(), EMPTY_BYTES));
}
return new ColumnQualifierFilter(skvi, colset);
} else {
SortedSet<Range> ranges = GraphuloUtil.d4mRowToRanges(colFilter);
assert ranges.size() > 0;
if (ranges.size() == 1) { // single range - use ColumnSliceFilter
Range r = ranges.first();
Map<String, String> map = new HashMap<>();
String start = r.isInfiniteStartKey() ? null : r.getStartKey().getRow().toString(),
end = r.isInfiniteStopKey() ? null : r.getEndKey().getRow().toString();
boolean startInclusive = true, endInclusive = true;
if (start != null) map.put(ColumnSliceFilter.START_BOUND, start);
if (end != null) map.put(ColumnSliceFilter.END_BOUND, end);
map.put(ColumnSliceFilter.START_INCLUSIVE, String.valueOf(startInclusive));
map.put(ColumnSliceFilter.END_INCLUSIVE, String.valueOf(endInclusive));
SortedKeyValueIterator<Key, Value> filter = new ColumnSliceFilter();
filter.init(skvi, map, env);
return filter;
} else { // multiple ranges
SortedKeyValueIterator<Key, Value> filter = new D4mRangeFilter();
filter.init(
skvi,
D4mRangeFilter.iteratorSetting(1, D4mRangeFilter.KeyPart.COLQ, colFilter).getOptions(),
env);
return filter;
}
}
}
@Override
public synchronized void init(
SortedKeyValueIterator<K, V> source, Map<String, String> options, IteratorEnvironment env)
throws IOException {
this.source = source;
source.init(source, options, env);
}
@Override
public void seek(Range range, Collection<ByteSequence> columnFamilies, boolean inclusive)
throws IOException {
if (!range.isInfiniteStartKey() || !range.isInfiniteStopKey())
log.warn("Range is not infinite: " + range);
source.seek(range, columnFamilies, inclusive);
IteratorAdapter ia = new IteratorAdapter(source);
SortedMap<Key, Value> map =
MemMatrixUtil.matrixToMap(
new TreeMap<Key, Value>(),
MemMatrixUtil.doInverse(MemMatrixUtil.buildMatrix(ia, matrixSize), numIterations));
// DebugUtil.printMapFull(map.entrySet().iterator());
mapIterator = new MapIterator(map);
mapIterator.init(null, null, null);
mapIterator.seek(range, columnFamilies, inclusive);
}
