arrow_ord/
partition.rs

Help
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
18//! Defines partition kernel for `ArrayRef`
19
20use std::ops::Range;
21
22use arrow_array::{Array, ArrayRef};
23use arrow_buffer::BooleanBuffer;
24use arrow_schema::ArrowError;
25
26use crate::cmp::distinct;
27use crate::sort::SortColumn;
28
29/// A computed set of partitions, see [`partition`]
30#[derive(Debug, Clone)]
31pub struct Partitions(Option<BooleanBuffer>);
32
33impl Partitions {
34    /// Returns the range of each partition
35    ///
36    /// Consecutive ranges will be contiguous: i.e [`(a, b)` and `(b, c)`], and
37    /// `start = 0` and `end = self.len()` for the first and last range respectively
38    pub fn ranges(&self) -> Vec<Range<usize>> {
39        let boundaries = match &self.0 {
40            Some(boundaries) => boundaries,
41            None => return vec![],
42        };
43
44        let mut out = vec![];
45        let mut current = 0;
46        for idx in boundaries.set_indices() {
47            let t = current;
48            current = idx + 1;
49            out.push(t..current)
50        }
51        let last = boundaries.len() + 1;
52        if current != last {
53            out.push(current..last)
54        }
55        out
56    }
57
58    /// Returns the number of partitions
59    pub fn len(&self) -> usize {
60        match &self.0 {
61            Some(b) => b.count_set_bits() + 1,
62            None => 0,
63        }
64    }
65
66    /// Returns true if this contains no partitions
67    pub fn is_empty(&self) -> bool {
68        self.0.is_none()
69    }
70}
71
72/// Given a list of lexicographically sorted columns, computes the [`Partitions`],
73/// where a partition consists of the set of consecutive rows with equal values
74///
75/// Returns an error if no columns are specified or all columns do not
76/// have the same number of rows.
77///
78/// # Example:
79///
80/// For example, given columns `x`, `y` and `z`, calling
81/// [`partition`]`(values, (x, y))` will divide the
82/// rows into ranges where the values of `(x, y)` are equal:
83///
84/// ```text
85/// ┌ ─ ┬───┬ ─ ─┌───┐─ ─ ┬───┬ ─ ─ ┐
86///     │ 1 │    │ 1 │    │ A │        Range: 0..1 (x=1, y=1)
87/// ├ ─ ┼───┼ ─ ─├───┤─ ─ ┼───┼ ─ ─ ┤
88///     │ 1 │    │ 2 │    │ B │
89/// │   ├───┤    ├───┤    ├───┤     │
90///     │ 1 │    │ 2 │    │ C │        Range: 1..4 (x=1, y=2)
91/// │   ├───┤    ├───┤    ├───┤     │
92///     │ 1 │    │ 2 │    │ D │
93/// ├ ─ ┼───┼ ─ ─├───┤─ ─ ┼───┼ ─ ─ ┤
94///     │ 2 │    │ 1 │    │ E │        Range: 4..5 (x=2, y=1)
95/// ├ ─ ┼───┼ ─ ─├───┤─ ─ ┼───┼ ─ ─ ┤
96///     │ 3 │    │ 1 │    │ F │        Range: 5..6 (x=3, y=1)
97/// └ ─ ┴───┴ ─ ─└───┘─ ─ ┴───┴ ─ ─ ┘
98///
99///       x        y        z     partition(&[x, y])
100/// ```
101///
102/// # Example Code
103///
104/// ```
105/// # use std::{sync::Arc, ops::Range};
106/// # use arrow_array::{RecordBatch, Int64Array, StringArray, ArrayRef};
107/// # use arrow_ord::sort::{SortColumn, SortOptions};
108/// # use arrow_ord::partition::partition;
109/// let batch = RecordBatch::try_from_iter(vec![
110///     ("x", Arc::new(Int64Array::from(vec![1, 1, 1, 1, 2, 3])) as ArrayRef),
111///     ("y", Arc::new(Int64Array::from(vec![1, 2, 2, 2, 1, 1])) as ArrayRef),
112///     ("z", Arc::new(StringArray::from(vec!["A", "B", "C", "D", "E", "F"])) as ArrayRef),
113/// ]).unwrap();
114///
115/// // Partition on first two columns
116/// let ranges = partition(&batch.columns()[..2]).unwrap().ranges();
117///
118/// let expected = vec![
119///     (0..1),
120///     (1..4),
121///     (4..5),
122///     (5..6),
123/// ];
124///
125/// assert_eq!(ranges, expected);
126/// ```
127pub fn partition(columns: &[ArrayRef]) -> Result<Partitions, ArrowError> {
128    if columns.is_empty() {
129        return Err(ArrowError::InvalidArgumentError(
130            "Partition requires at least one column".to_string(),
131        ));
132    }
133    let num_rows = columns[0].len();
134    if columns.iter().any(|item| item.len() != num_rows) {
135        return Err(ArrowError::InvalidArgumentError(
136            "Partition columns have different row counts".to_string(),
137        ));
138    };
139
140    match num_rows {
141        0 => return Ok(Partitions(None)),
142        1 => return Ok(Partitions(Some(BooleanBuffer::new_unset(0)))),
143        _ => {}
144    }
145
146    let acc = find_boundaries(&columns[0])?;
147    let acc = columns
148        .iter()
149        .skip(1)
150        .try_fold(acc, |acc, c| find_boundaries(c.as_ref()).map(|b| &acc | &b))?;
151
152    Ok(Partitions(Some(acc)))
153}
154
155/// Returns a mask with bits set whenever the value or nullability changes
156fn find_boundaries(v: &dyn Array) -> Result<BooleanBuffer, ArrowError> {
157    let slice_len = v.len() - 1;
158    let v1 = v.slice(0, slice_len);
159    let v2 = v.slice(1, slice_len);
160    Ok(distinct(&v1, &v2)?.values().clone())
161}
162
163/// Use [`partition`] instead. Given a list of already sorted columns, find
164/// partition ranges that would partition lexicographically equal values across
165/// columns.
166///
167/// The returned vec would be of size k where k is cardinality of the sorted values; Consecutive
168/// values will be connected: (a, b) and (b, c), where start = 0 and end = n for the first and last
169/// range.
170#[deprecated(note = "Use partition")]
171pub fn lexicographical_partition_ranges(
172    columns: &[SortColumn],
173) -> Result<impl Iterator<Item = Range<usize>> + '_, ArrowError> {
174    let cols: Vec<_> = columns.iter().map(|x| x.values.clone()).collect();
175    Ok(partition(&cols)?.ranges().into_iter())
176}
177
178#[cfg(test)]
179mod tests {
180    use std::sync::Arc;
181
182    use arrow_array::*;
183    use arrow_schema::DataType;
184
185    use super::*;
186
187    #[test]
188    fn test_partition_empty() {
189        let err = partition(&[]).unwrap_err();
190        assert_eq!(
191            err.to_string(),
192            "Invalid argument error: Partition requires at least one column"
193        );
194    }
195
196    #[test]
197    fn test_partition_unaligned_rows() {
198        let input = vec![
199            Arc::new(Int64Array::from(vec![None, Some(-1)])) as _,
200            Arc::new(StringArray::from(vec![Some("foo")])) as _,
201        ];
202        let err = partition(&input).unwrap_err();
203        assert_eq!(
204            err.to_string(),
205            "Invalid argument error: Partition columns have different row counts"
206        )
207    }
208
209    #[test]
210    fn test_partition_small() {
211        let results = partition(&[
212            Arc::new(Int32Array::new(vec![].into(), None)) as _,
213            Arc::new(Int32Array::new(vec![].into(), None)) as _,
214            Arc::new(Int32Array::new(vec![].into(), None)) as _,
215        ])
216        .unwrap();
217        assert_eq!(results.len(), 0);
218        assert!(results.is_empty());
219
220        let results = partition(&[
221            Arc::new(Int32Array::from(vec![1])) as _,
222            Arc::new(Int32Array::from(vec![1])) as _,
223        ])
224        .unwrap()
225        .ranges();
226        assert_eq!(results.len(), 1);
227        assert_eq!(results[0], 0..1);
228    }
229
230    #[test]
231    fn test_partition_single_column() {
232        let a = Int64Array::from(vec![1, 2, 2, 2, 2, 2, 2, 2, 9]);
233        let input = vec![Arc::new(a) as _];
234        assert_eq!(
235            partition(&input).unwrap().ranges(),
236            vec![(0..1), (1..8), (8..9)],
237        );
238    }
239
240    #[test]
241    fn test_partition_all_equal_values() {
242        let a = Int64Array::from_value(1, 1000);
243        let input = vec![Arc::new(a) as _];
244        assert_eq!(partition(&input).unwrap().ranges(), vec![(0..1000)]);
245    }
246
247    #[test]
248    fn test_partition_all_null_values() {
249        let input = vec![
250            new_null_array(&DataType::Int8, 1000),
251            new_null_array(&DataType::UInt16, 1000),
252        ];
253        assert_eq!(partition(&input).unwrap().ranges(), vec![(0..1000)]);
254    }
255
256    #[test]
257    fn test_partition_unique_column_1() {
258        let input = vec![
259            Arc::new(Int64Array::from(vec![None, Some(-1)])) as _,
260            Arc::new(StringArray::from(vec![Some("foo"), Some("bar")])) as _,
261        ];
262        assert_eq!(partition(&input).unwrap().ranges(), vec![(0..1), (1..2)],);
263    }
264
265    #[test]
266    fn test_partition_unique_column_2() {
267        let input = vec![
268            Arc::new(Int64Array::from(vec![None, Some(-1), Some(-1)])) as _,
269            Arc::new(StringArray::from(vec![
270                Some("foo"),
271                Some("bar"),
272                Some("apple"),
273            ])) as _,
274        ];
275        assert_eq!(
276            partition(&input).unwrap().ranges(),
277            vec![(0..1), (1..2), (2..3),],
278        );
279    }
280
281    #[test]
282    fn test_partition_non_unique_column_1() {
283        let input = vec![
284            Arc::new(Int64Array::from(vec![None, Some(-1), Some(-1), Some(1)])) as _,
285            Arc::new(StringArray::from(vec![
286                Some("foo"),
287                Some("bar"),
288                Some("bar"),
289                Some("bar"),
290            ])) as _,
291        ];
292        assert_eq!(
293            partition(&input).unwrap().ranges(),
294            vec![(0..1), (1..3), (3..4),],
295        );
296    }
297
298    #[test]
299    fn test_partition_masked_nulls() {
300        let input = vec![
301            Arc::new(Int64Array::new(vec![1; 9].into(), None)) as _,
302            Arc::new(Int64Array::new(
303                vec![1, 1, 2, 2, 2, 3, 3, 3, 3].into(),
304                Some(vec![false, true, true, true, true, false, false, true, false].into()),
305            )) as _,
306            Arc::new(Int64Array::new(
307                vec![1, 1, 2, 2, 2, 2, 2, 3, 7].into(),
308                Some(vec![true, true, true, true, false, true, true, true, false].into()),
309            )) as _,
310        ];
311
312        assert_eq!(
313            partition(&input).unwrap().ranges(),
314            vec![(0..1), (1..2), (2..4), (4..5), (5..7), (7..8), (8..9)],
315        );
316    }
317}
arrow_ord/partition.rs

arrow_ord/
partition.rs
