[Enhancement] make LIKE operations' statistics consistent (#55023)

Signed-off-by: Elmi Ahmadov <[email protected]> (cherry picked from commit 6085037)
StarRocks · Feb 7, 2025 · b079868 · b079868
1 parent 360fcef
commit b079868
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 0 deletions.
diff --git a/...e/src/main/java/com/starrocks/sql/optimizer/statistics/ExpressionStatisticCalculator.java b/...e/src/main/java/com/starrocks/sql/optimizer/statistics/ExpressionStatisticCalculator.java
@@ -613,6 +613,12 @@ private ColumnStatistic binaryExpressionCalculate(CallOperator callOperator, Col
                     averageRowSize = Math.max(1, left.getAverageRowSize() - right.getAverageRowSize());
                     distinctValues = left.getDistinctValuesCount() * averageRowSize / left.getAverageRowSize();
                     break;
+                case FunctionSet.LIKE:
+                case FunctionSet.ILIKE:
+                    minValue = 0;
+                    maxValue = 1;
+                    distinctValues = 2;
+                    break;
                 default:
                     return ColumnStatistic.unknown();
             }

diff --git a/.../test/java/com/starrocks/sql/optimizer/statistics/ExpressionStatisticsCalculatorTest.java b/.../test/java/com/starrocks/sql/optimizer/statistics/ExpressionStatisticsCalculatorTest.java
@@ -506,6 +506,18 @@ public void testBinaryFunctionCall() {
         columnStatistic = ExpressionStatisticCalculator.calculate(callOperator, builder.build());
         Assert.assertEquals(-1, columnStatistic.getMinValue(), 0.001);
         Assert.assertEquals(1, columnStatistic.getMaxValue(), 0.001);
+
+        callOperator = new CallOperator(FunctionSet.LIKE, Type.BOOLEAN, Lists.newArrayList(left, right));
+        columnStatistic = ExpressionStatisticCalculator.calculate(callOperator, builder.build());
+        Assert.assertEquals(0, columnStatistic.getMinValue(), 0.001);
+        Assert.assertEquals(1, columnStatistic.getMaxValue(), 0.001);
+        Assert.assertEquals(2, columnStatistic.getDistinctValuesCount(), 0.001);
+
+        callOperator = new CallOperator(FunctionSet.ILIKE, Type.BOOLEAN, Lists.newArrayList(left, right));
+        columnStatistic = ExpressionStatisticCalculator.calculate(callOperator, builder.build());
+        Assert.assertEquals(0, columnStatistic.getMinValue(), 0.001);
+        Assert.assertEquals(1, columnStatistic.getMaxValue(), 0.001);
+        Assert.assertEquals(2, columnStatistic.getDistinctValuesCount(), 0.001);
         // test multiply/divide column rang is negative
         builder = Statistics.builder();
         leftStatistic = new ColumnStatistic(-100, -10, 0, 0, 20);