arrow_array/array/
fixed_size_list_array.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 crate::array::print_long_array;
19use crate::builder::{FixedSizeListBuilder, PrimitiveBuilder};
20use crate::iterator::FixedSizeListIter;
21use crate::{Array, ArrayAccessor, ArrayRef, ArrowPrimitiveType, make_array};
22use arrow_buffer::ArrowNativeType;
23use arrow_buffer::buffer::NullBuffer;
24use arrow_data::{ArrayData, ArrayDataBuilder};
25use arrow_schema::{ArrowError, DataType, FieldRef};
26use std::any::Any;
27use std::sync::Arc;
28
29/// An array of [fixed length lists], similar to JSON arrays
30/// (e.g. `["A", "B"]`).
31///
32/// Lists are represented using a `values` child
33/// array where each list has a fixed size of `value_length`.
34///
35/// Use [`FixedSizeListBuilder`] to construct a [`FixedSizeListArray`].
36///
37/// # Representation
38///
39/// A [`FixedSizeListArray`] can represent a list of values of any other
40/// supported Arrow type. Each element of the `FixedSizeListArray` itself is
41/// a list which may contain NULL and non-null values,
42/// or may itself be NULL.
43///
44/// For example, this `FixedSizeListArray` stores lists of strings:
45///
46/// ```text
47/// ┌─────────────┐
48/// │    [A,B]    │
49/// ├─────────────┤
50/// │    NULL     │
51/// ├─────────────┤
52/// │   [C,NULL]  │
53/// └─────────────┘
54/// ```
55///
56/// The `values` of this `FixedSizeListArray`s are stored in a child
57/// [`StringArray`] where logical null values take up `values_length` slots in the array
58/// as shown in the following diagram. The logical values
59/// are shown on the left, and the actual `FixedSizeListArray` encoding on the right
60///
61/// ```text
62///                                 ┌ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ┐
63///                                                         ┌ ─ ─ ─ ─ ─ ─ ─ ─┐
64///  ┌─────────────┐                │     ┌───┐               ┌───┐ ┌──────┐      │
65///  │   [A,B]     │                      │ 1 │             │ │ 1 │ │  A   │ │ 0
66///  ├─────────────┤                │     ├───┤               ├───┤ ├──────┤      │
67///  │    NULL     │                      │ 0 │             │ │ 1 │ │  B   │ │ 1
68///  ├─────────────┤                │     ├───┤               ├───┤ ├──────┤      │
69///  │  [C,NULL]   │                      │ 1 │             │ │ 0 │ │ ???? │ │ 2
70///  └─────────────┘                │     └───┘               ├───┤ ├──────┤      │
71///                                                         | │ 0 │ │ ???? │ │ 3
72///  Logical Values                 │   Validity              ├───┤ ├──────┤      │
73///                                     (nulls)             │ │ 1 │ │  C   │ │ 4
74///                                 │                         ├───┤ ├──────┤      │
75///                                                         │ │ 0 │ │ ???? │ │ 5
76///                                 │                         └───┘ └──────┘      │
77///                                                         │     Values     │
78///                                 │   FixedSizeListArray        (Array)         │
79///                                                         └ ─ ─ ─ ─ ─ ─ ─ ─┘
80///                                 └ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ┘
81/// ```
82///
83/// # Example
84///
85/// ```
86/// # use std::sync::Arc;
87/// # use arrow_array::{Array, FixedSizeListArray, Int32Array};
88/// # use arrow_data::ArrayData;
89/// # use arrow_schema::{DataType, Field};
90/// # use arrow_buffer::Buffer;
91/// // Construct a value array
92/// let value_data = ArrayData::builder(DataType::Int32)
93///     .len(9)
94///     .add_buffer(Buffer::from_slice_ref(&[0, 1, 2, 3, 4, 5, 6, 7, 8]))
95///     .build()
96///     .unwrap();
97/// let list_data_type = DataType::FixedSizeList(
98///     Arc::new(Field::new_list_field(DataType::Int32, false)),
99///     3,
100/// );
101/// let list_data = ArrayData::builder(list_data_type.clone())
102///     .len(3)
103///     .add_child_data(value_data.clone())
104///     .build()
105///     .unwrap();
106/// let list_array = FixedSizeListArray::from(list_data);
107/// let list0 = list_array.value(0);
108/// let list1 = list_array.value(1);
109/// let list2 = list_array.value(2);
110///
111/// assert_eq!( &[0, 1, 2], list0.as_any().downcast_ref::<Int32Array>().unwrap().values());
112/// assert_eq!( &[3, 4, 5], list1.as_any().downcast_ref::<Int32Array>().unwrap().values());
113/// assert_eq!( &[6, 7, 8], list2.as_any().downcast_ref::<Int32Array>().unwrap().values());
114/// ```
115///
116/// [`StringArray`]: crate::array::StringArray
117/// [fixed length lists]: https://arrow.apache.org/docs/format/Columnar.html#fixed-size-list-layout
118#[derive(Clone)]
119pub struct FixedSizeListArray {
120    data_type: DataType, // Must be DataType::FixedSizeList(value_length)
121    values: ArrayRef,
122    nulls: Option<NullBuffer>,
123    value_length: i32,
124    len: usize,
125}
126
127impl FixedSizeListArray {
128    /// Create a new [`FixedSizeListArray`] with `size` element size, panicking on failure.
129    ///
130    /// Note that if `size == 0` and `nulls` is `None` (a degenerate, non-nullable
131    /// `FixedSizeListArray`), this function will set the length of the array to 0.
132    ///
133    /// If you would like to have a degenerate, non-nullable `FixedSizeListArray` with arbitrary
134    /// length, use the [`try_new_with_length()`] constructor.
135    ///
136    /// [`try_new_with_length()`]: Self::try_new_with_length
137    ///
138    /// # Panics
139    ///
140    /// Panics if [`Self::try_new`] returns an error
141    pub fn new(field: FieldRef, size: i32, values: ArrayRef, nulls: Option<NullBuffer>) -> Self {
142        Self::try_new(field, size, values, nulls).unwrap()
143    }
144
145    /// Create a new [`FixedSizeListArray`] from the provided parts without validation.
146    ///
147    /// # Safety
148    /// - `size >= 0`
149    /// - `values.len() == len * size as usize`
150    /// - `nulls.len() == len` if `nulls` is `Some`
151    /// - `field.data_type() == values.data_type()`
152    pub unsafe fn new_unchecked(
153        field: FieldRef,
154        size: i32,
155        values: ArrayRef,
156        nulls: Option<NullBuffer>,
157        len: usize,
158    ) -> Self {
159        if cfg!(feature = "force_validate") {
160            return Self::try_new_with_length(field, size, values, nulls, len).unwrap();
161        }
162        Self {
163            data_type: DataType::FixedSizeList(field, size),
164            values,
165            value_length: size,
166            nulls,
167            len,
168        }
169    }
170
171    /// Create a new [`FixedSizeListArray`] from the provided parts, returning an error on failure.
172    ///
173    /// Note that if `size == 0` and `nulls` is `None` (a degenerate, non-nullable
174    /// `FixedSizeListArray`), this function will set the length of the array to 0.
175    ///
176    /// If you would like to have a degenerate, non-nullable `FixedSizeListArray` with arbitrary
177    /// length, use the [`try_new_with_length()`] constructor.
178    ///
179    /// [`try_new_with_length()`]: Self::try_new_with_length
180    ///
181    /// # Errors
182    ///
183    /// * `size < 0`
184    /// * `values.len() != nulls.len() * size` if `nulls` is `Some`
185    /// * `values.data_type() != field.data_type()`
186    /// * `!field.is_nullable() && !nulls.expand(size).contains(values.logical_nulls())`
187    pub fn try_new(
188        field: FieldRef,
189        size: i32,
190        values: ArrayRef,
191        nulls: Option<NullBuffer>,
192    ) -> Result<Self, ArrowError> {
193        let s = size.to_usize().ok_or_else(|| {
194            ArrowError::InvalidArgumentError(format!("Size cannot be negative, got {size}"))
195        })?;
196
197        if s == 0 {
198            // Note that for degenerate (`size == 0`) and non-nullable `FixedSizeList`s, we will set
199            // the length to 0 (`_or_default`).
200            let len = nulls.as_ref().map(|x| x.len()).unwrap_or_default();
201
202            Self::try_new_with_length(field, size, values, nulls, len)
203        } else {
204            if values.len() % s != 0 {
205                return Err(ArrowError::InvalidArgumentError(format!(
206                    "Incorrect length of values buffer for FixedSizeListArray, \
207                     expected a multiple of {s} got {}",
208                    values.len(),
209                )));
210            }
211
212            let len = values.len() / s;
213
214            // Check that the null buffer length is correct (if it exists).
215            if let Some(null_buffer) = &nulls {
216                if s * null_buffer.len() != values.len() {
217                    return Err(ArrowError::InvalidArgumentError(format!(
218                        "Incorrect length of values buffer for FixedSizeListArray, \
219                            expected {} got {}",
220                        s * null_buffer.len(),
221                        values.len(),
222                    )));
223                }
224            }
225
226            Self::try_new_with_length(field, size, values, nulls, len)
227        }
228    }
229
230    /// Create a new [`FixedSizeListArray`] from the provided parts, returning an error on failure.
231    ///
232    /// This method exists to allow the construction of arbitrary length degenerate (`size == 0`)
233    /// and non-nullable `FixedSizeListArray`s. If you want a nullable `FixedSizeListArray`, then
234    /// you can use [`try_new()`] instead.
235    ///
236    /// [`try_new()`]: Self::try_new
237    ///
238    /// # Errors
239    ///
240    /// * `size < 0`
241    /// * `nulls.len() != len` if `nulls` is `Some`
242    /// * `values.len() != len * size`
243    /// * `values.data_type() != field.data_type()`
244    /// * `!field.is_nullable() && !nulls.expand(size).contains(values.logical_nulls())`
245    pub fn try_new_with_length(
246        field: FieldRef,
247        size: i32,
248        values: ArrayRef,
249        nulls: Option<NullBuffer>,
250        len: usize,
251    ) -> Result<Self, ArrowError> {
252        let s = size.to_usize().ok_or_else(|| {
253            ArrowError::InvalidArgumentError(format!("Size cannot be negative, got {size}"))
254        })?;
255
256        if let Some(null_buffer) = &nulls {
257            if null_buffer.len() != len {
258                return Err(ArrowError::InvalidArgumentError(format!(
259                    "Invalid null buffer for FixedSizeListArray, expected {len} found {}",
260                    null_buffer.len()
261                )));
262            }
263        }
264
265        if s == 0 && !values.is_empty() {
266            return Err(ArrowError::InvalidArgumentError(format!(
267                "An degenerate FixedSizeListArray should have no underlying values, found {} values",
268                values.len()
269            )));
270        }
271
272        if values.len() != len * s {
273            return Err(ArrowError::InvalidArgumentError(format!(
274                "Incorrect length of values buffer for FixedSizeListArray, expected {} got {}",
275                len * s,
276                values.len(),
277            )));
278        }
279
280        if field.data_type() != values.data_type() {
281            return Err(ArrowError::InvalidArgumentError(format!(
282                "FixedSizeListArray expected data type {} got {} for {:?}",
283                field.data_type(),
284                values.data_type(),
285                field.name()
286            )));
287        }
288
289        if let Some(a) = values.logical_nulls() {
290            let nulls_valid = field.is_nullable()
291                || nulls
292                    .as_ref()
293                    .map(|n| n.expand(size as _).contains(&a))
294                    .unwrap_or_default()
295                || (nulls.is_none() && a.null_count() == 0);
296
297            if !nulls_valid {
298                return Err(ArrowError::InvalidArgumentError(format!(
299                    "Found unmasked nulls for non-nullable FixedSizeListArray field {:?}",
300                    field.name()
301                )));
302            }
303        }
304
305        let data_type = DataType::FixedSizeList(field, size);
306        Ok(Self {
307            data_type,
308            values,
309            value_length: size,
310            nulls,
311            len,
312        })
313    }
314
315    /// Create a new [`FixedSizeListArray`] of length `len` where all values are null
316    ///
317    /// # Panics
318    ///
319    /// Panics if
320    ///
321    /// * `size < 0`
322    /// * `size * len` would overflow `usize`
323    pub fn new_null(field: FieldRef, size: i32, len: usize) -> Self {
324        let capacity = size.to_usize().unwrap().checked_mul(len).unwrap();
325        Self {
326            values: make_array(ArrayData::new_null(field.data_type(), capacity)),
327            data_type: DataType::FixedSizeList(field, size),
328            nulls: Some(NullBuffer::new_null(len)),
329            value_length: size,
330            len,
331        }
332    }
333
334    /// Deconstruct this array into its constituent parts
335    pub fn into_parts(self) -> (FieldRef, i32, ArrayRef, Option<NullBuffer>) {
336        let f = match self.data_type {
337            DataType::FixedSizeList(f, _) => f,
338            _ => unreachable!(),
339        };
340        (f, self.value_length, self.values, self.nulls)
341    }
342
343    /// Returns a reference to the values of this list.
344    pub fn values(&self) -> &ArrayRef {
345        &self.values
346    }
347
348    /// Returns a clone of the value type of this list.
349    pub fn value_type(&self) -> DataType {
350        self.values.data_type().clone()
351    }
352
353    /// Returns ith value of this list array.
354    ///
355    /// Note: This method does not check for nulls and the value is arbitrary
356    /// (but still well-defined) if [`is_null`](Self::is_null) returns true for the index.
357    ///
358    /// # Panics
359    /// Panics if index `i` is out of bounds
360    pub fn value(&self, i: usize) -> ArrayRef {
361        self.values
362            .slice(self.value_offset_at(i), self.value_length() as usize)
363    }
364
365    /// Returns the offset for value at index `i`.
366    ///
367    /// Note this doesn't do any bound checking, for performance reason.
368    #[inline]
369    pub fn value_offset(&self, i: usize) -> i32 {
370        self.value_offset_at(i) as i32
371    }
372
373    /// Returns the length for an element.
374    ///
375    /// All elements have the same length as the array is a fixed size.
376    #[inline]
377    pub const fn value_length(&self) -> i32 {
378        self.value_length
379    }
380
381    #[inline]
382    const fn value_offset_at(&self, i: usize) -> usize {
383        i * self.value_length as usize
384    }
385
386    /// Returns a zero-copy slice of this array with the indicated offset and length.
387    pub fn slice(&self, offset: usize, len: usize) -> Self {
388        assert!(
389            offset.saturating_add(len) <= self.len,
390            "the length + offset of the sliced FixedSizeListArray cannot exceed the existing length"
391        );
392        let size = self.value_length as usize;
393
394        Self {
395            data_type: self.data_type.clone(),
396            values: self.values.slice(offset * size, len * size),
397            nulls: self.nulls.as_ref().map(|n| n.slice(offset, len)),
398            value_length: self.value_length,
399            len,
400        }
401    }
402
403    /// Creates a [`FixedSizeListArray`] from an iterator of primitive values
404    /// # Example
405    /// ```
406    /// # use arrow_array::FixedSizeListArray;
407    /// # use arrow_array::types::Int32Type;
408    ///
409    /// let data = vec![
410    ///    Some(vec![Some(0), Some(1), Some(2)]),
411    ///    None,
412    ///    Some(vec![Some(3), None, Some(5)]),
413    ///    Some(vec![Some(6), Some(7), Some(45)]),
414    /// ];
415    /// let list_array = FixedSizeListArray::from_iter_primitive::<Int32Type, _, _>(data, 3);
416    /// println!("{:?}", list_array);
417    /// ```
418    pub fn from_iter_primitive<T, P, I>(iter: I, length: i32) -> Self
419    where
420        T: ArrowPrimitiveType,
421        P: IntoIterator<Item = Option<<T as ArrowPrimitiveType>::Native>>,
422        I: IntoIterator<Item = Option<P>>,
423    {
424        let l = length as usize;
425        let iter = iter.into_iter();
426        let size_hint = iter.size_hint().0;
427        let mut builder = FixedSizeListBuilder::with_capacity(
428            PrimitiveBuilder::<T>::with_capacity(size_hint * l),
429            length,
430            size_hint,
431        );
432
433        for i in iter {
434            match i {
435                Some(p) => {
436                    for t in p {
437                        builder.values().append_option(t);
438                    }
439                    builder.append(true);
440                }
441                None => {
442                    builder.values().append_nulls(l);
443                    builder.append(false)
444                }
445            }
446        }
447        builder.finish()
448    }
449
450    /// constructs a new iterator
451    pub fn iter(&self) -> FixedSizeListIter<'_> {
452        FixedSizeListIter::new(self)
453    }
454}
455
456impl From<ArrayData> for FixedSizeListArray {
457    fn from(data: ArrayData) -> Self {
458        let (data_type, len, nulls, offset, _buffers, child_data) = data.into_parts();
459
460        let value_length = match data_type {
461            DataType::FixedSizeList(_, len) => len,
462            data_type => {
463                panic!(
464                    "FixedSizeListArray data should contain a FixedSizeList data type, got {data_type}"
465                )
466            }
467        };
468
469        let size = value_length as usize;
470        let values = make_array(child_data[0].slice(offset * size, len * size));
471        Self {
472            data_type,
473            values,
474            nulls,
475            value_length,
476            len,
477        }
478    }
479}
480
481impl From<FixedSizeListArray> for ArrayData {
482    fn from(array: FixedSizeListArray) -> Self {
483        let builder = ArrayDataBuilder::new(array.data_type)
484            .len(array.len)
485            .nulls(array.nulls)
486            .child_data(vec![array.values.to_data()]);
487
488        unsafe { builder.build_unchecked() }
489    }
490}
491
492/// SAFETY: Correctly implements the contract of Arrow Arrays
493unsafe impl Array for FixedSizeListArray {
494    fn as_any(&self) -> &dyn Any {
495        self
496    }
497
498    fn to_data(&self) -> ArrayData {
499        self.clone().into()
500    }
501
502    fn into_data(self) -> ArrayData {
503        self.into()
504    }
505
506    fn data_type(&self) -> &DataType {
507        &self.data_type
508    }
509
510    fn slice(&self, offset: usize, length: usize) -> ArrayRef {
511        Arc::new(self.slice(offset, length))
512    }
513
514    fn len(&self) -> usize {
515        self.len
516    }
517
518    fn is_empty(&self) -> bool {
519        self.len == 0
520    }
521
522    fn shrink_to_fit(&mut self) {
523        self.values.shrink_to_fit();
524        if let Some(nulls) = &mut self.nulls {
525            nulls.shrink_to_fit();
526        }
527    }
528
529    fn offset(&self) -> usize {
530        0
531    }
532
533    fn nulls(&self) -> Option<&NullBuffer> {
534        self.nulls.as_ref()
535    }
536
537    fn logical_null_count(&self) -> usize {
538        // More efficient that the default implementation
539        self.null_count()
540    }
541
542    fn get_buffer_memory_size(&self) -> usize {
543        let mut size = self.values.get_buffer_memory_size();
544        if let Some(n) = self.nulls.as_ref() {
545            size += n.buffer().capacity();
546        }
547        size
548    }
549
550    fn get_array_memory_size(&self) -> usize {
551        let mut size = std::mem::size_of::<Self>() + self.values.get_array_memory_size();
552        if let Some(n) = self.nulls.as_ref() {
553            size += n.buffer().capacity();
554        }
555        size
556    }
557
558    #[cfg(feature = "pool")]
559    fn claim(&self, pool: &dyn arrow_buffer::MemoryPool) {
560        self.values.claim(pool);
561        if let Some(nulls) = &self.nulls {
562            nulls.claim(pool);
563        }
564    }
565}
566
567impl super::ListLikeArray for FixedSizeListArray {
568    fn values(&self) -> &ArrayRef {
569        self.values()
570    }
571
572    fn element_range(&self, index: usize) -> std::ops::Range<usize> {
573        let value_length = self.value_length().as_usize();
574        let offset = index * value_length;
575        offset..(offset + value_length)
576    }
577}
578
579impl ArrayAccessor for FixedSizeListArray {
580    type Item = ArrayRef;
581
582    fn value(&self, index: usize) -> Self::Item {
583        FixedSizeListArray::value(self, index)
584    }
585
586    unsafe fn value_unchecked(&self, index: usize) -> Self::Item {
587        FixedSizeListArray::value(self, index)
588    }
589}
590
591impl std::fmt::Debug for FixedSizeListArray {
592    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
593        write!(f, "FixedSizeListArray<{}>\n[\n", self.value_length())?;
594        print_long_array(self, f, |array, index, f| {
595            std::fmt::Debug::fmt(&array.value(index), f)
596        })?;
597        write!(f, "]")
598    }
599}
600
601impl ArrayAccessor for &FixedSizeListArray {
602    type Item = ArrayRef;
603
604    fn value(&self, index: usize) -> Self::Item {
605        FixedSizeListArray::value(self, index)
606    }
607
608    unsafe fn value_unchecked(&self, index: usize) -> Self::Item {
609        FixedSizeListArray::value(self, index)
610    }
611}
612
613#[cfg(test)]
614mod tests {
615    use arrow_buffer::{BooleanBuffer, Buffer, bit_util};
616    use arrow_schema::Field;
617
618    use crate::cast::AsArray;
619    use crate::types::Int32Type;
620    use crate::{Int32Array, new_empty_array};
621
622    use super::*;
623
624    #[test]
625    fn test_fixed_size_list_array() {
626        // Construct a value array
627        let value_data = ArrayData::builder(DataType::Int32)
628            .len(9)
629            .add_buffer(Buffer::from_slice_ref([0, 1, 2, 3, 4, 5, 6, 7, 8]))
630            .build()
631            .unwrap();
632
633        // Construct a list array from the above two
634        let list_data_type =
635            DataType::FixedSizeList(Arc::new(Field::new_list_field(DataType::Int32, false)), 3);
636        let list_data = ArrayData::builder(list_data_type.clone())
637            .len(3)
638            .add_child_data(value_data.clone())
639            .build()
640            .unwrap();
641        let list_array = FixedSizeListArray::from(list_data);
642
643        assert_eq!(value_data, list_array.values().to_data());
644        assert_eq!(DataType::Int32, list_array.value_type());
645        assert_eq!(3, list_array.len());
646        assert_eq!(0, list_array.null_count());
647        assert_eq!(6, list_array.value_offset(2));
648        assert_eq!(3, list_array.value_length());
649        assert_eq!(0, list_array.value(0).as_primitive::<Int32Type>().value(0));
650        for i in 0..3 {
651            assert!(list_array.is_valid(i));
652            assert!(!list_array.is_null(i));
653        }
654
655        // Now test with a non-zero offset
656        let list_data = ArrayData::builder(list_data_type)
657            .len(2)
658            .offset(1)
659            .add_child_data(value_data.clone())
660            .build()
661            .unwrap();
662        let list_array = FixedSizeListArray::from(list_data);
663
664        assert_eq!(value_data.slice(3, 6), list_array.values().to_data());
665        assert_eq!(DataType::Int32, list_array.value_type());
666        assert_eq!(2, list_array.len());
667        assert_eq!(0, list_array.null_count());
668        assert_eq!(3, list_array.value(0).as_primitive::<Int32Type>().value(0));
669        assert_eq!(3, list_array.value_offset(1));
670        assert_eq!(3, list_array.value_length());
671    }
672
673    #[test]
674    #[should_panic(expected = "assertion failed: end <= self.len()")]
675    // Different error messages, so skip for now
676    // https://github.com/apache/arrow-rs/issues/1545
677    #[cfg(not(feature = "force_validate"))]
678    fn test_fixed_size_list_array_unequal_children() {
679        // Construct a value array
680        let value_data = ArrayData::builder(DataType::Int32)
681            .len(8)
682            .add_buffer(Buffer::from_slice_ref([0, 1, 2, 3, 4, 5, 6, 7]))
683            .build()
684            .unwrap();
685
686        // Construct a list array from the above two
687        let list_data_type =
688            DataType::FixedSizeList(Arc::new(Field::new_list_field(DataType::Int32, false)), 3);
689        let list_data = unsafe {
690            ArrayData::builder(list_data_type)
691                .len(3)
692                .add_child_data(value_data)
693                .build_unchecked()
694        };
695        drop(FixedSizeListArray::from(list_data));
696    }
697
698    #[test]
699    fn test_fixed_size_list_array_slice() {
700        // Construct a value array
701        let value_data = ArrayData::builder(DataType::Int32)
702            .len(10)
703            .add_buffer(Buffer::from_slice_ref([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]))
704            .build()
705            .unwrap();
706
707        // Set null buts for the nested array:
708        //  [[0, 1], null, null, [6, 7], [8, 9]]
709        // 01011001 00000001
710        let mut null_bits: [u8; 1] = [0; 1];
711        bit_util::set_bit(&mut null_bits, 0);
712        bit_util::set_bit(&mut null_bits, 3);
713        bit_util::set_bit(&mut null_bits, 4);
714
715        // Construct a fixed size list array from the above two
716        let list_data_type =
717            DataType::FixedSizeList(Arc::new(Field::new_list_field(DataType::Int32, false)), 2);
718        let list_data = ArrayData::builder(list_data_type)
719            .len(5)
720            .add_child_data(value_data.clone())
721            .null_bit_buffer(Some(Buffer::from(null_bits)))
722            .build()
723            .unwrap();
724        let list_array = FixedSizeListArray::from(list_data);
725
726        assert_eq!(value_data, list_array.values().to_data());
727        assert_eq!(DataType::Int32, list_array.value_type());
728        assert_eq!(5, list_array.len());
729        assert_eq!(2, list_array.null_count());
730        assert_eq!(6, list_array.value_offset(3));
731        assert_eq!(2, list_array.value_length());
732
733        let sliced_array = list_array.slice(1, 4);
734        assert_eq!(4, sliced_array.len());
735        assert_eq!(2, sliced_array.null_count());
736
737        for i in 0..sliced_array.len() {
738            if bit_util::get_bit(&null_bits, 1 + i) {
739                assert!(sliced_array.is_valid(i));
740            } else {
741                assert!(sliced_array.is_null(i));
742            }
743        }
744
745        // Check offset and length for each non-null value.
746        let sliced_list_array = sliced_array
747            .as_any()
748            .downcast_ref::<FixedSizeListArray>()
749            .unwrap();
750        assert_eq!(2, sliced_list_array.value_length());
751        assert_eq!(4, sliced_list_array.value_offset(2));
752        assert_eq!(6, sliced_list_array.value_offset(3));
753    }
754
755    #[test]
756    #[should_panic(expected = "the offset of the new Buffer cannot exceed the existing length")]
757    fn test_fixed_size_list_array_index_out_of_bound() {
758        // Construct a value array
759        let value_data = ArrayData::builder(DataType::Int32)
760            .len(10)
761            .add_buffer(Buffer::from_slice_ref([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]))
762            .build()
763            .unwrap();
764
765        // Set null buts for the nested array:
766        //  [[0, 1], null, null, [6, 7], [8, 9]]
767        // 01011001 00000001
768        let mut null_bits: [u8; 1] = [0; 1];
769        bit_util::set_bit(&mut null_bits, 0);
770        bit_util::set_bit(&mut null_bits, 3);
771        bit_util::set_bit(&mut null_bits, 4);
772
773        // Construct a fixed size list array from the above two
774        let list_data_type =
775            DataType::FixedSizeList(Arc::new(Field::new_list_field(DataType::Int32, false)), 2);
776        let list_data = ArrayData::builder(list_data_type)
777            .len(5)
778            .add_child_data(value_data)
779            .null_bit_buffer(Some(Buffer::from(null_bits)))
780            .build()
781            .unwrap();
782        let list_array = FixedSizeListArray::from(list_data);
783
784        list_array.value(10);
785    }
786
787    #[test]
788    fn test_fixed_size_list_constructors() {
789        let values = Arc::new(Int32Array::from_iter([
790            Some(1),
791            Some(2),
792            None,
793            None,
794            Some(3),
795            Some(4),
796        ]));
797
798        let field = Arc::new(Field::new_list_field(DataType::Int32, true));
799        let list = FixedSizeListArray::new(field.clone(), 2, values.clone(), None);
800        assert_eq!(list.len(), 3);
801
802        let nulls = NullBuffer::new_null(3);
803        let list = FixedSizeListArray::new(field.clone(), 2, values.clone(), Some(nulls));
804        assert_eq!(list.len(), 3);
805
806        let list = FixedSizeListArray::new(field.clone(), 3, values.clone(), None);
807        assert_eq!(list.len(), 2);
808
809        let err = FixedSizeListArray::try_new(field.clone(), 4, values.clone(), None).unwrap_err();
810        assert_eq!(
811            err.to_string(),
812            "Invalid argument error: Incorrect length of values buffer for FixedSizeListArray, \
813             expected a multiple of 4 got 6",
814        );
815
816        let err =
817            FixedSizeListArray::try_new_with_length(field.clone(), 4, values.clone(), None, 1)
818                .unwrap_err();
819        assert_eq!(
820            err.to_string(),
821            "Invalid argument error: Incorrect length of values buffer for FixedSizeListArray, expected 4 got 6"
822        );
823
824        let err = FixedSizeListArray::try_new(field.clone(), -1, values.clone(), None).unwrap_err();
825        assert_eq!(
826            err.to_string(),
827            "Invalid argument error: Size cannot be negative, got -1"
828        );
829
830        let nulls = NullBuffer::new_null(2);
831        let err = FixedSizeListArray::try_new(field, 2, values.clone(), Some(nulls)).unwrap_err();
832        assert_eq!(
833            err.to_string(),
834            "Invalid argument error: Incorrect length of values buffer for FixedSizeListArray, expected 4 got 6"
835        );
836
837        let field = Arc::new(Field::new_list_field(DataType::Int32, false));
838        let err = FixedSizeListArray::try_new(field.clone(), 2, values.clone(), None).unwrap_err();
839        assert_eq!(
840            err.to_string(),
841            "Invalid argument error: Found unmasked nulls for non-nullable FixedSizeListArray field \"item\""
842        );
843
844        // Valid as nulls in child masked by parent
845        let nulls = NullBuffer::new(BooleanBuffer::new(Buffer::from([0b0000101]), 0, 3));
846        FixedSizeListArray::new(field, 2, values.clone(), Some(nulls));
847
848        let field = Arc::new(Field::new_list_field(DataType::Int64, true));
849        let err = FixedSizeListArray::try_new(field, 2, values, None).unwrap_err();
850        assert_eq!(
851            err.to_string(),
852            "Invalid argument error: FixedSizeListArray expected data type Int64 got Int32 for \"item\""
853        );
854    }
855
856    #[test]
857    fn degenerate_fixed_size_list() {
858        let field = Arc::new(Field::new_list_field(DataType::Int32, true));
859        let nulls = NullBuffer::new_null(2);
860        let values = new_empty_array(&DataType::Int32);
861        let list = FixedSizeListArray::new(field.clone(), 0, values.clone(), Some(nulls.clone()));
862        assert_eq!(list.len(), 2);
863
864        // Test invalid null buffer length.
865        let err = FixedSizeListArray::try_new_with_length(
866            field.clone(),
867            0,
868            values.clone(),
869            Some(nulls),
870            5,
871        )
872        .unwrap_err();
873        assert_eq!(
874            err.to_string(),
875            "Invalid argument error: Invalid null buffer for FixedSizeListArray, expected 5 found 2"
876        );
877
878        // Test non-empty values for degenerate list.
879        let non_empty_values = Arc::new(Int32Array::from(vec![1, 2, 3]));
880        let err =
881            FixedSizeListArray::try_new_with_length(field.clone(), 0, non_empty_values, None, 3)
882                .unwrap_err();
883        assert_eq!(
884            err.to_string(),
885            "Invalid argument error: An degenerate FixedSizeListArray should have no underlying values, found 3 values"
886        );
887    }
888
889    #[test]
890    fn test_fixed_size_list_new_null_len() {
891        let field = Arc::new(Field::new_list_field(DataType::Int32, true));
892        let array = FixedSizeListArray::new_null(field, 2, 5);
893        assert_eq!(array.len(), 5);
894    }
895}
arrow_array/array/fixed_size_list_array.rs

arrow_array/array/
fixed_size_list_array.rs
