datafusion_functions_nested/
flatten.rs

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
// 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.

//! [`ScalarUDFImpl`] definitions for flatten function.

use crate::utils::make_scalar_function;
use arrow_array::{ArrayRef, GenericListArray, OffsetSizeTrait};
use arrow_buffer::OffsetBuffer;
use arrow_schema::DataType;
use arrow_schema::DataType::{FixedSizeList, LargeList, List, Null};
use datafusion_common::cast::{
    as_generic_list_array, as_large_list_array, as_list_array,
};
use datafusion_common::{exec_err, Result};
use datafusion_expr::scalar_doc_sections::DOC_SECTION_ARRAY;
use datafusion_expr::{
    ColumnarValue, Documentation, ScalarUDFImpl, Signature, Volatility,
};
use std::any::Any;
use std::sync::{Arc, OnceLock};

make_udf_expr_and_func!(
    Flatten,
    flatten,
    array,
    "flattens an array of arrays into a single array.",
    flatten_udf
);

#[derive(Debug)]
pub(super) struct Flatten {
    signature: Signature,
    aliases: Vec<String>,
}
impl Flatten {
    pub fn new() -> Self {
        Self {
            signature: Signature::array(Volatility::Immutable),
            aliases: vec![],
        }
    }
}

impl ScalarUDFImpl for Flatten {
    fn as_any(&self) -> &dyn Any {
        self
    }

    fn name(&self) -> &str {
        "flatten"
    }

    fn signature(&self) -> &Signature {
        &self.signature
    }

    fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
        fn get_base_type(data_type: &DataType) -> Result<DataType> {
            match data_type {
                List(field) | FixedSizeList(field, _)
                    if matches!(field.data_type(), List(_) | FixedSizeList(_, _)) =>
                {
                    get_base_type(field.data_type())
                }
                LargeList(field) if matches!(field.data_type(), LargeList(_)) => {
                    get_base_type(field.data_type())
                }
                Null | List(_) | LargeList(_) => Ok(data_type.to_owned()),
                FixedSizeList(field, _) => Ok(List(Arc::clone(field))),
                _ => exec_err!(
                    "Not reachable, data_type should be List, LargeList or FixedSizeList"
                ),
            }
        }

        let data_type = get_base_type(&arg_types[0])?;
        Ok(data_type)
    }

    fn invoke(&self, args: &[ColumnarValue]) -> Result<ColumnarValue> {
        make_scalar_function(flatten_inner)(args)
    }

    fn aliases(&self) -> &[String] {
        &self.aliases
    }

    fn documentation(&self) -> Option<&Documentation> {
        Some(get_flatten_doc())
    }
}
static DOCUMENTATION: OnceLock<Documentation> = OnceLock::new();

fn get_flatten_doc() -> &'static Documentation {
    DOCUMENTATION.get_or_init(|| {
        Documentation::builder()
            .with_doc_section(DOC_SECTION_ARRAY)
            .with_description(
                "Converts an array of arrays to a flat array.\n\n- Applies to any depth of nested arrays\n- Does not change arrays that are already flat\n\nThe flattened array contains all the elements from all source arrays.",
            )
            .with_syntax_example("flatten(array)")
            .with_sql_example(
                r#"```sql
> select flatten([[1, 2], [3, 4]]);
+------------------------------+
| flatten(List([1,2], [3,4]))  |
+------------------------------+
| [1, 2, 3, 4]                 |
+------------------------------+
```"#,
            )
            .with_argument(
                "array",
                "Array expression. Can be a constant, column, or function, and any combination of array operators.",
            )
            .build()
            .unwrap()
    })
}

/// Flatten SQL function
pub fn flatten_inner(args: &[ArrayRef]) -> Result<ArrayRef> {
    if args.len() != 1 {
        return exec_err!("flatten expects one argument");
    }

    let array_type = args[0].data_type();
    match array_type {
        List(_) => {
            let list_arr = as_list_array(&args[0])?;
            let flattened_array = flatten_internal::<i32>(list_arr.clone(), None)?;
            Ok(Arc::new(flattened_array) as ArrayRef)
        }
        LargeList(_) => {
            let list_arr = as_large_list_array(&args[0])?;
            let flattened_array = flatten_internal::<i64>(list_arr.clone(), None)?;
            Ok(Arc::new(flattened_array) as ArrayRef)
        }
        Null => Ok(Arc::clone(&args[0])),
        _ => {
            exec_err!("flatten does not support type '{array_type:?}'")
        }
    }
}

fn flatten_internal<O: OffsetSizeTrait>(
    list_arr: GenericListArray<O>,
    indexes: Option<OffsetBuffer<O>>,
) -> Result<GenericListArray<O>> {
    let (field, offsets, values, _) = list_arr.clone().into_parts();
    let data_type = field.data_type();

    match data_type {
        // Recursively get the base offsets for flattened array
        List(_) | LargeList(_) => {
            let sub_list = as_generic_list_array::<O>(&values)?;
            if let Some(indexes) = indexes {
                let offsets = get_offsets_for_flatten(offsets, indexes);
                flatten_internal::<O>(sub_list.clone(), Some(offsets))
            } else {
                flatten_internal::<O>(sub_list.clone(), Some(offsets))
            }
        }
        // Reach the base level, create a new list array
        _ => {
            if let Some(indexes) = indexes {
                let offsets = get_offsets_for_flatten(offsets, indexes);
                let list_arr = GenericListArray::<O>::new(field, offsets, values, None);
                Ok(list_arr)
            } else {
                Ok(list_arr)
            }
        }
    }
}

// Create new offsets that are equivalent to `flatten` the array.
fn get_offsets_for_flatten<O: OffsetSizeTrait>(
    offsets: OffsetBuffer<O>,
    indexes: OffsetBuffer<O>,
) -> OffsetBuffer<O> {
    let buffer = offsets.into_inner();
    let offsets: Vec<O> = indexes
        .iter()
        .map(|i| buffer[i.to_usize().unwrap()])
        .collect();
    OffsetBuffer::new(offsets.into())
}