datafusion_spark/function/aggregate/
collect.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18use arrow::array::ArrayRef;
19use arrow::datatypes::{DataType, Field, FieldRef};
20use datafusion_common::utils::SingleRowListArrayBuilder;
21use datafusion_common::{Result, ScalarValue};
22use datafusion_expr::function::{AccumulatorArgs, StateFieldsArgs};
23use datafusion_expr::utils::format_state_name;
24use datafusion_expr::{Accumulator, AggregateUDFImpl, Signature, Volatility};
25use datafusion_functions_aggregate::array_agg::{
26    ArrayAggAccumulator, DistinctArrayAggAccumulator,
27};
28use std::sync::Arc;
29
30// Spark implementation of collect_list/collect_set aggregate function.
31// Differs from DataFusion ArrayAgg in the following ways:
32// - ignores NULL inputs
33// - returns an empty list when all inputs are NULL
34// - does not support ordering
35
36// <https://spark.apache.org/docs/latest/api/sql/index.html#collect_list>
37#[derive(Debug, PartialEq, Eq, Hash)]
38pub struct SparkCollectList {
39    signature: Signature,
40}
41
42impl Default for SparkCollectList {
43    fn default() -> Self {
44        Self::new()
45    }
46}
47
48impl SparkCollectList {
49    pub fn new() -> Self {
50        Self {
51            signature: Signature::any(1, Volatility::Immutable),
52        }
53    }
54}
55
56impl AggregateUDFImpl for SparkCollectList {
57    fn name(&self) -> &str {
58        "collect_list"
59    }
60
61    fn signature(&self) -> &Signature {
62        &self.signature
63    }
64
65    fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
66        Ok(DataType::List(Arc::new(Field::new_list_field(
67            arg_types[0].clone(),
68            true,
69        ))))
70    }
71
72    fn state_fields(&self, args: StateFieldsArgs) -> Result<Vec<FieldRef>> {
73        Ok(vec![
74            Field::new_list(
75                format_state_name(args.name, "collect_list"),
76                Field::new_list_field(args.input_fields[0].data_type().clone(), true),
77                true,
78            )
79            .into(),
80        ])
81    }
82
83    fn accumulator(&self, acc_args: AccumulatorArgs) -> Result<Box<dyn Accumulator>> {
84        let field = &acc_args.expr_fields[0];
85        let data_type = field.data_type().clone();
86        let ignore_nulls = true;
87        Ok(Box::new(NullToEmptyListAccumulator::new(
88            ArrayAggAccumulator::try_new(&data_type, ignore_nulls)?,
89            data_type,
90        )))
91    }
92}
93
94// <https://spark.apache.org/docs/latest/api/sql/index.html#collect_set>
95#[derive(Debug, PartialEq, Eq, Hash)]
96pub struct SparkCollectSet {
97    signature: Signature,
98}
99
100impl Default for SparkCollectSet {
101    fn default() -> Self {
102        Self::new()
103    }
104}
105
106impl SparkCollectSet {
107    pub fn new() -> Self {
108        Self {
109            signature: Signature::any(1, Volatility::Immutable),
110        }
111    }
112}
113
114impl AggregateUDFImpl for SparkCollectSet {
115    fn name(&self) -> &str {
116        "collect_set"
117    }
118
119    fn signature(&self) -> &Signature {
120        &self.signature
121    }
122
123    fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
124        Ok(DataType::List(Arc::new(Field::new_list_field(
125            arg_types[0].clone(),
126            true,
127        ))))
128    }
129
130    fn state_fields(&self, args: StateFieldsArgs) -> Result<Vec<FieldRef>> {
131        Ok(vec![
132            Field::new_list(
133                format_state_name(args.name, "collect_set"),
134                Field::new_list_field(args.input_fields[0].data_type().clone(), true),
135                true,
136            )
137            .into(),
138        ])
139    }
140
141    fn accumulator(&self, acc_args: AccumulatorArgs) -> Result<Box<dyn Accumulator>> {
142        let field = &acc_args.expr_fields[0];
143        let data_type = field.data_type().clone();
144        let ignore_nulls = true;
145        Ok(Box::new(NullToEmptyListAccumulator::new(
146            DistinctArrayAggAccumulator::try_new(&data_type, None, ignore_nulls)?,
147            data_type,
148        )))
149    }
150}
151
152/// Wrapper accumulator that returns an empty list instead of NULL when all inputs are NULL.
153/// This implements Spark's behavior for collect_list and collect_set.
154#[derive(Debug)]
155struct NullToEmptyListAccumulator<T: Accumulator> {
156    inner: T,
157    data_type: DataType,
158}
159
160impl<T: Accumulator> NullToEmptyListAccumulator<T> {
161    pub fn new(inner: T, data_type: DataType) -> Self {
162        Self { inner, data_type }
163    }
164}
165
166impl<T: Accumulator> Accumulator for NullToEmptyListAccumulator<T> {
167    fn update_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
168        self.inner.update_batch(values)
169    }
170
171    fn merge_batch(&mut self, states: &[ArrayRef]) -> Result<()> {
172        self.inner.merge_batch(states)
173    }
174
175    fn state(&mut self) -> Result<Vec<ScalarValue>> {
176        self.inner.state()
177    }
178
179    fn evaluate(&mut self) -> Result<ScalarValue> {
180        let result = self.inner.evaluate()?;
181        if result.is_null() {
182            let empty_array = arrow::array::new_empty_array(&self.data_type);
183            Ok(SingleRowListArrayBuilder::new(empty_array).build_list_scalar())
184        } else {
185            Ok(result)
186        }
187    }
188
189    fn size(&self) -> usize {
190        self.inner.size() + self.data_type.size()
191    }
192}
datafusion_spark/function/aggregate/collect.rs

datafusion_spark/function/aggregate/
collect.rs
