[GitHub] [arrow] jhorstmann commented on a change in pull request #8222: ARROW-10043: [Rust][DataFusion] Implement COUNT(DISTINCT col)

[GitBox <gi...@apache.org>]

jhorstmann commented on a change in pull request #8222:
URL: https://github.com/apache/arrow/pull/8222#discussion_r491869035



##########
File path: rust/datafusion/src/physical_plan/distinct_expressions.rs
##########
@@ -0,0 +1,303 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Implementations for DISTINCT expressions, e.g. `COUNT(DISTINCT c)`
+
+use std::cell::RefCell;
+use std::fmt::Debug;
+use std::hash::Hash;
+use std::rc::Rc;
+use std::sync::Arc;
+
+use arrow::array::ArrayRef;
+use arrow::array::{
+    Int16Array, Int32Array, Int64Array, Int8Array, PrimitiveArrayOps, UInt16Array,
+    UInt32Array, UInt64Array, UInt8Array,
+};
+use arrow::datatypes::{DataType, Schema};
+use arrow::record_batch::RecordBatch;
+
+use fnv::FnvHashSet;
+
+use crate::error::{ExecutionError, Result};
+use crate::logical_plan::ScalarValue;
+use crate::physical_plan::expressions::Column;
+use crate::physical_plan::hash_aggregate::AggregateMode;
+use crate::physical_plan::{Accumulator, AggregateExpr, PhysicalExpr};
+
+/// Enumeration of types that can be accumulated into a distinct set of values.
+#[derive(Debug, PartialEq, Eq, Hash, Clone)]
+enum DistinctScalarValue {
+    Int8(i8),
+    Int16(i16),
+    Int32(i32),
+    Int64(i64),
+    UInt8(u8),
+    UInt16(u16),
+    UInt32(u32),
+    UInt64(u64),
+}
+
+/// For a given expression, maps its Arrow DataType into a LargeList of the
+/// same DataType.
+fn list_of(expr: Arc<dyn PhysicalExpr>, input_schema: &Schema) -> Result<DataType> {
+    let value_data_type = expr.data_type(input_schema)?;
+
+    match value_data_type {
+        DataType::Int8
+        | DataType::Int16
+        | DataType::Int32
+        | DataType::Int64
+        | DataType::UInt8
+        | DataType::UInt16
+        | DataType::UInt32
+        | DataType::UInt64 => Ok(DataType::LargeList(Box::new(value_data_type))),
+        _ => Err(ExecutionError::NotImplemented(
+            "Unsupported column data type for DISTINCT".to_string(),
+        )),
+    }
+}
+
+fn accumulate_scalar(
+    accum: &mut FnvHashSet<DistinctScalarValue>,
+    value: Option<ScalarValue>,
+) -> Result<()> {
+    let accum_value = match value {
+        Some(ScalarValue::Int8(v)) => Some(DistinctScalarValue::Int8(v)),
+        Some(ScalarValue::Int16(v)) => Some(DistinctScalarValue::Int16(v)),
+        Some(ScalarValue::Int32(v)) => Some(DistinctScalarValue::Int32(v)),
+        Some(ScalarValue::Int64(v)) => Some(DistinctScalarValue::Int64(v)),
+        Some(ScalarValue::UInt8(v)) => Some(DistinctScalarValue::UInt8(v)),
+        Some(ScalarValue::UInt16(v)) => Some(DistinctScalarValue::UInt16(v)),
+        Some(ScalarValue::UInt32(v)) => Some(DistinctScalarValue::UInt32(v)),
+        Some(ScalarValue::UInt64(v)) => Some(DistinctScalarValue::UInt64(v)),
+        Some(ScalarValue::Null) => None,
+        _ => {
+            return Err(ExecutionError::NotImplemented(
+                "Unsupported scalar value for DISTINCT accumulator".to_string(),
+            ))
+        }
+    };
+
+    match accum_value {
+        Some(v) => {
+            accum.insert(v);
+        }
+        None => {}
+    }
+
+    Ok(())
+}
+
+macro_rules! accum_batch {
+    ($ARRAY_TY:ident, $DISTINCT_SCALAR_TY: path, $ARRAY: expr, $ACCUM: expr) => {{
+        let array = $ARRAY.as_any().downcast_ref::<$ARRAY_TY>().ok_or_else(|| {
+            ExecutionError::ExecutionError("Error downcasting array".to_string())
+        })?;
+
+        for i in 0..array.len() {
+            $ACCUM.insert($DISTINCT_SCALAR_TY(array.value(i)));
+        }
+
+        Ok(())
+    }};
+}
+
+#[derive(Debug)]
+struct DistinctValuesAccumulator {
+    values: FnvHashSet<DistinctScalarValue>,

Review comment:
       A more efficient implementation for distinct integers could be a bitset




----------------------------------------------------------------
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:
users@infra.apache.org