ajantha-bhat commented on a change in pull request #3616:
[CARBONDATA-3548]Polygon expression processing using unknown expression and
filtering performance improvement
URL: https://github.com/apache/carbondata/pull/3616#discussion_r378275250
##########
File path:
geo/src/main/java/org/apache/carbondata/geo/scan/expression/PolygonExpression.java
##########
@@ -41,64 +41,85 @@
* InExpression with list of all the IDs present in those list of ranges.
*/
@InterfaceAudience.Internal
-public class PolygonExpression extends Expression {
+public class PolygonExpression extends UnknownExpression implements
ConditionalExpression {
private String polygon;
- private String columnName;
private CustomIndex<List<Long[]>> handler;
- private List<Expression> children = new ArrayList<Expression>();
+ private List<Long[]> ranges = new ArrayList<Long[]>();
+ private ColumnExpression column;
+ private ExpressionResult trueExpRes;
+ private ExpressionResult falseExpRes;
public PolygonExpression(String polygon, String columnName, CustomIndex
handler) {
this.polygon = polygon;
this.handler = handler;
- this.columnName = columnName;
+ this.column = new ColumnExpression(columnName, DataTypes.LONG);
+ this.trueExpRes = new ExpressionResult(DataTypes.BOOLEAN, true);
+ this.falseExpRes = new ExpressionResult(DataTypes.BOOLEAN, false);
}
- private void buildExpression(List<Long[]> ranges) {
- // Build InExpression with list of all the values present in the ranges
- List<Expression> inList = new ArrayList<Expression>();
+ private void validate(List<Long[]> ranges) {
+ // Validate the ranges
for (Long[] range : ranges) {
if (range.length != 2) {
throw new RuntimeException("Handler query must return list of ranges
with each range "
+ "containing minimum and maximum values");
}
- for (long i = range[0]; i <= range[1]; i++) {
- inList.add(new LiteralExpression(i, DataTypes.LONG));
- }
}
- children.add(new InExpression(new ColumnExpression(columnName,
DataTypes.LONG),
- new ListExpression(inList)));
}
/**
- * This method builds InExpression with list of all the values present in
the list of ranges of
- * IDs.
+ * This method calls the query processor and gets the list of ranges of IDs.
*/
private void processExpression() {
- List<Long[]> ranges;
try {
ranges = handler.query(polygon);
+ validate(ranges);
} catch (Exception e) {
throw new RuntimeException(e);
}
- buildExpression(ranges);
+ }
+
+ private boolean rangeBinarySearch(List<Long[]> ranges, long searchForNumber)
{
+ Long[] range;
+ int low = 0, mid, high = ranges.size() - 1;
+ while (low <= high) {
+ mid = low + ((high - low) / 2);
+ range = ranges.get(mid);
+ if (searchForNumber >= range[0]) {
+ if (searchForNumber <= range[1]) {
+ // Return true if the number is between min and max values of the
range
+ return true;
+ } else {
+ // Number is bigger than this range's min and max. Search on the
right side of the range
+ low = mid + 1;
+ }
+ } else {
+ // Number is smaller than this range's min and max. Search on the left
side of the range
+ high = mid - 1;
+ }
+ }
+ return false;
}
@Override
public ExpressionResult evaluate(RowIntf value) {
- throw new UnsupportedOperationException("Operation not supported for
Polygon expression");
+ if (rangeBinarySearch(ranges, (Long) value.getVal(0))) {
Review comment:
Data is already sorted. So may be in future can take advantage of it ,
instead of checking each row exist in a range.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
