From 6d8c7ad9361021c66330c95e8f137e661efa18b7 Mon Sep 17 00:00:00 2001
From: zhuqi-lucas <821684824@qq.com>
Date: Thu, 13 Mar 2025 17:53:25 +0800
Subject: [PATCH 1/2] feat: support ApproxDistinct with utf8view

---
 .../src/approx_distinct.rs                    | 40 ++++++++++++++++++-
 .../test_files/aggregate_skip_partial.slt     | 21 ++++++++++
 2 files changed, 60 insertions(+), 1 deletion(-)

diff --git a/datafusion/functions-aggregate/src/approx_distinct.rs b/datafusion/functions-aggregate/src/approx_distinct.rs
index 1d378fff176f..82eeefba19a8 100644
--- a/datafusion/functions-aggregate/src/approx_distinct.rs
+++ b/datafusion/functions-aggregate/src/approx_distinct.rs
@@ -18,7 +18,7 @@
 //! Defines physical expressions that can evaluated at runtime during query execution
 
 use crate::hyperloglog::HyperLogLog;
-use arrow::array::BinaryArray;
+use arrow::array::{BinaryArray, StringViewArray};
 use arrow::array::{
     GenericBinaryArray, GenericStringArray, OffsetSizeTrait, PrimitiveArray,
 };
@@ -126,6 +126,28 @@ where
     }
 }
 
+#[derive(Debug)]
+struct StringViewHLLAccumulator<T>
+where
+    T: OffsetSizeTrait,
+{
+    hll: HyperLogLog<String>,
+    phantom_data: PhantomData<T>,
+}
+
+impl<T> StringViewHLLAccumulator<T>
+where
+    T: OffsetSizeTrait,
+{
+    /// new approx_distinct accumulator
+    pub fn new() -> Self {
+        Self {
+            hll: HyperLogLog::new(),
+            phantom_data: PhantomData,
+        }
+    }
+}
+
 #[derive(Debug)]
 struct BinaryHLLAccumulator<T>
 where
@@ -197,6 +219,21 @@ where
     default_accumulator_impl!();
 }
 
+impl<T> Accumulator for StringViewHLLAccumulator<T>
+where
+    T: OffsetSizeTrait,
+{
+    fn update_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
+        let array: &StringViewArray = downcast_value!(values[0], StringViewArray);
+        // flatten because we would skip nulls
+        self.hll
+            .extend(array.iter().flatten().map(|s| s.to_string()));
+        Ok(())
+    }
+
+    default_accumulator_impl!();
+}
+
 impl<T> Accumulator for StringHLLAccumulator<T>
 where
     T: OffsetSizeTrait,
@@ -311,6 +348,7 @@ impl AggregateUDFImpl for ApproxDistinct {
             DataType::Int64 => Box::new(NumericHLLAccumulator::<Int64Type>::new()),
             DataType::Utf8 => Box::new(StringHLLAccumulator::<i32>::new()),
             DataType::LargeUtf8 => Box::new(StringHLLAccumulator::<i64>::new()),
+            DataType::Utf8View => Box::new(StringViewHLLAccumulator::<i32>::new()),
             DataType::Binary => Box::new(BinaryHLLAccumulator::<i32>::new()),
             DataType::LargeBinary => Box::new(BinaryHLLAccumulator::<i64>::new()),
             other => {
diff --git a/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt b/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
index 3a4d641abf68..163e27f435cb 100644
--- a/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
+++ b/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
@@ -298,6 +298,27 @@ SELECT c2, approx_distinct(c1), approx_distinct(c5) FROM aggregate_test_100 GROU
 4 5 23
 5 5 14
 
+# Test approx_distinct for varchar(with Utf8View) / int
+statement ok
+CREATE TABLE aggregate_test_100_utf8view AS SELECT
+      arrow_cast(c1, 'Utf8View') as c1,
+      c2,
+      c5
+FROM aggregate_test_100;
+
+# Test approx_distinct for varchar / int
+query III
+SELECT c2, approx_distinct(c1), approx_distinct(c5) FROM aggregate_test_100_utf8view GROUP BY c2 ORDER BY c2;
+----
+1 5 22
+2 5 22
+3 5 19
+4 5 23
+5 5 14
+
+statement ok
+DROP TABLE aggregate_test_100_utf8view;
+
 # Test count with nullable fields
 query III
 SELECT c2, count(c3), count(c11) FROM aggregate_test_100_null GROUP BY c2 ORDER BY c2;

From 473151cf512047914802968cc4f919d7bf0b65f5 Mon Sep 17 00:00:00 2001
From: zhuqi-lucas <821684824@qq.com>
Date: Thu, 13 Mar 2025 22:58:50 +0800
Subject: [PATCH 2/2] Address comment

---
 datafusion/functions-aggregate/src/approx_distinct.rs         | 1 -
 datafusion/sqllogictest/test_files/aggregate_skip_partial.slt | 2 +-
 2 files changed, 1 insertion(+), 2 deletions(-)

diff --git a/datafusion/functions-aggregate/src/approx_distinct.rs b/datafusion/functions-aggregate/src/approx_distinct.rs
index 82eeefba19a8..c97dba1925ca 100644
--- a/datafusion/functions-aggregate/src/approx_distinct.rs
+++ b/datafusion/functions-aggregate/src/approx_distinct.rs
@@ -139,7 +139,6 @@ impl<T> StringViewHLLAccumulator<T>
 where
     T: OffsetSizeTrait,
 {
-    /// new approx_distinct accumulator
     pub fn new() -> Self {
         Self {
             hll: HyperLogLog::new(),
diff --git a/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt b/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
index 163e27f435cb..8755918cd16c 100644
--- a/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
+++ b/datafusion/sqllogictest/test_files/aggregate_skip_partial.slt
@@ -306,7 +306,7 @@ CREATE TABLE aggregate_test_100_utf8view AS SELECT
       c5
 FROM aggregate_test_100;
 
-# Test approx_distinct for varchar / int
+# Test approx_distinct for varchar(with Utf8View) / int
 query III
 SELECT c2, approx_distinct(c1), approx_distinct(c5) FROM aggregate_test_100_utf8view GROUP BY c2 ORDER BY c2;
 ----