Apache DataFusion SQL Query Engine
fix: preserve state in DistinctMedianAccumulator::evaluate() for window frame queries (#19887)
Kumar Ujjawal6b1a1cb81eb4
diff --git a/datafusion/functions-aggregate/src/median.rs b/datafusion/functions-aggregate/src/median.rsindex f137ae0801..57cb14d7c1 100644--- a/datafusion/functions-aggregate/src/median.rs+++ b/datafusion/functions-aggregate/src/median.rs@@ -566,10 +566,8 @@ impl<T: ArrowNumericType + Debug> Accumulator for DistinctMedianAccumulator<T> {}fn evaluate(&mut self) -> Result<ScalarValue> {- let mut d = std::mem::take(&mut self.distinct_values.values)- .into_iter()- .map(|v| v.0)- .collect::<Vec<_>>();+ let mut d: Vec<T::Native> =+ self.distinct_values.values.iter().map(|v| v.0).collect();let median = calculate_median::<T>(&mut d);ScalarValue::new_primitive::<T>(median, &self.data_type)}diff --git a/datafusion/sqllogictest/test_files/aggregate.slt b/datafusion/sqllogictest/test_files/aggregate.sltindex a5f3ef0413..be24178637 100644--- a/datafusion/sqllogictest/test_files/aggregate.slt+++ b/datafusion/sqllogictest/test_files/aggregate.slt@@ -1226,6 +1226,32 @@ ORDER BY tags, timestamp;4 tag2 90 67.5 82.55 tag2 100 70 90++# Test distinct median non-sliding window+query ITRR+SELECT+ timestamp,+ tags,+ value,+ median(DISTINCT value) OVER (+ PARTITION BY tags+ ORDER BY timestamp+ ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW+ ) AS distinct_median+FROM median_window_test+ORDER BY tags, timestamp;+----+1 tag1 10 10+2 tag1 20 15+3 tag1 30 20+4 tag1 40 25+5 tag1 50 30+1 tag2 60 60+2 tag2 70 65+3 tag2 80 70+4 tag2 90 75+5 tag2 100 80+statement okDROP TABLE median_window_test;