fix(query): fix and check total_buffer_len and total_bytes_len

src/common/arrow/src/arrow/array/binview/mod.rs

@@ -166,6 +166,18 @@ impl<T: ViewType + ?Sized> BinaryViewArrayGeneric<T> {
         total_bytes_len: usize,
         total_buffer_len: usize,
     ) -> Self {
+        #[cfg(debug_assertions)]
+        {
+            if total_bytes_len != UNKNOWN_LEN as usize {
+                let total = views.iter().map(|v| v.length as usize).sum::<usize>();
+                assert_eq!(total, total_bytes_len);
+            }
+
+            if total_buffer_len != UNKNOWN_LEN as usize {
+                let total = buffers.iter().map(|v| v.len()).sum::<usize>();
+                assert_eq!(total, total_buffer_len);
+            }
+        }
         // # Safety
         // The caller must ensure
         // - the data is valid utf8 (if required)

src/common/arrow/src/arrow/array/binview/mutable.rs

@@ -263,12 +263,23 @@ impl<T: ViewType + ?Sized> MutableBinaryViewArray<T> {
         // Push and pop to get the properly encoded value.
         // For long string this leads to a dictionary encoding,
         // as we push the string only once in the buffers
+
+        let old_bytes_len = self.total_bytes_len;
+        let old_buffer_len = self.total_buffer_len;
+
         let view_value = value
             .map(|v| {
                 self.push_value_ignore_validity(v);
                 self.views.pop().unwrap()
             })
             .unwrap_or_default();
+
+        self.total_bytes_len +=
+            (self.total_bytes_len - old_bytes_len) * additional.saturating_sub(1);
+
+        self.total_buffer_len +=
+            (self.total_buffer_len - old_buffer_len) * additional.saturating_sub(1);
+
         self.views
             .extend(std::iter::repeat(view_value).take(additional));
     }

src/common/arrow/src/arrow/array/growable/binview.rs

@@ -74,7 +74,6 @@ impl<'a, T: ViewType + ?Sized> GrowableBinaryViewArray<'a, T> {
         capacity: usize,
     ) -> Self {
         let data_type = arrays[0].data_type().clone();
-
         // if any of the arrays has nulls, insertions from any array requires setting bits
         // as there is at least one array with nulls.
         if !use_validity & arrays.iter().any(|array| array.null_count() > 0) {
@@ -91,11 +90,8 @@ impl<'a, T: ViewType + ?Sized> GrowableBinaryViewArray<'a, T> {
                     .map(|buf| BufferKey { inner: buf })
             })
             .collect::<ArrowIndexSet<_>>();
-        let total_buffer_len = arrays
-            .iter()
-            .map(|arr| arr.data_buffers().len())
-            .sum::<usize>();
 
+        let total_buffer_len = buffers.iter().map(|v| v.inner.len()).sum();
         Self {
             arrays,
             data_type,

src/query/expression/src/kernels/filter.rs

@@ -358,7 +358,7 @@ impl<'a> FilterVisitor<'a> {
                         new_views,
                         values.data.data_buffers().clone(),
                         None,
-                        Some(values.data.total_buffer_len()),
+                        None,
                     )
                 };
                 StringColumn::new(new_col)

src/query/expression/src/kernels/take.rs

@@ -292,7 +292,7 @@ where I: databend_common_arrow::arrow::types::Index
                     new_views,
                     col.data.data_buffers().clone(),
                     None,
-                    Some(col.data.total_buffer_len()),
+                    None,
                 )
             };
             StringColumn::new(new_col)

src/query/expression/src/kernels/take_compact.rs

@@ -238,7 +238,7 @@ impl<'a> TakeCompactVisitor<'a> {
                 new_views,
                 col.data.data_buffers().clone(),
                 None,
-                Some(col.data.total_buffer_len()),
+                None,
             )
         };
         StringColumn::new(new_col)

src/query/expression/src/kernels/take_ranges.rs

@@ -239,7 +239,7 @@ impl<'a> TakeRangeVisitor<'a> {
                 new_views,
                 col.data.data_buffers().clone(),
                 None,
-                Some(col.data.total_buffer_len()),
+                None,
             )
         };
         StringColumn::new(new_col)

tests/sqllogictests/suites/query/join/left_outer.test

@@ -298,8 +298,18 @@ SELECT * FROM t1 LEFT JOIN t2 ON t1.a = t2.a;
 4 5 NULL NULL
 5 6 NULL NULL
 
+
+statement ok
+create or replace table t1 (a string) as select number from numbers(100);
+
+statement ok
+create or replace table t2 (a string) as select number from numbers(100);
+
+## just check it works or not
 statement ok
-set max_block_size = 65536;
+select * from (
+   select 'SN0LL' as k  from t1
+) as a1  left  join (select * from t2) as a2 on a1.k = a2.a;
 
 statement ok
 DROP TABLE IF EXISTS t1;