arrow_array/array/
byte_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::{get_offsets_from_buffer, print_long_array};
19use crate::builder::GenericByteBuilder;
20use crate::iterator::ArrayIter;
21use crate::types::ByteArrayType;
22use crate::types::bytes::ByteArrayNativeType;
23use crate::{Array, ArrayAccessor, ArrayRef, OffsetSizeTrait, Scalar};
24use arrow_buffer::{ArrowNativeType, Buffer, MutableBuffer};
25use arrow_buffer::{NullBuffer, OffsetBuffer};
26use arrow_data::{ArrayData, ArrayDataBuilder};
27use arrow_schema::{ArrowError, DataType};
28use std::any::Any;
29use std::sync::Arc;
30
31/// An array of [variable length byte arrays](https://arrow.apache.org/docs/format/Columnar.html#variable-size-binary-layout)
32///
33/// See [`StringArray`] and [`LargeStringArray`] for storing utf8 encoded string data
34///
35/// See [`BinaryArray`] and [`LargeBinaryArray`] for storing arbitrary bytes
36///
37/// # Example: From a Vec
38///
39/// ```
40/// # use arrow_array::{Array, GenericByteArray, types::Utf8Type};
41/// let arr: GenericByteArray<Utf8Type> = vec!["hello", "world", ""].into();
42/// assert_eq!(arr.value_data(), b"helloworld");
43/// assert_eq!(arr.value_offsets(), &[0, 5, 10, 10]);
44/// let values: Vec<_> = arr.iter().collect();
45/// assert_eq!(values, &[Some("hello"), Some("world"), Some("")]);
46/// ```
47///
48/// # Example: From an optional Vec
49///
50/// ```
51/// # use arrow_array::{Array, GenericByteArray, types::Utf8Type};
52/// let arr: GenericByteArray<Utf8Type> = vec![Some("hello"), Some("world"), Some(""), None].into();
53/// assert_eq!(arr.value_data(), b"helloworld");
54/// assert_eq!(arr.value_offsets(), &[0, 5, 10, 10, 10]);
55/// let values: Vec<_> = arr.iter().collect();
56/// assert_eq!(values, &[Some("hello"), Some("world"), Some(""), None]);
57/// ```
58///
59/// # Example: From an iterator of option
60///
61/// ```
62/// # use arrow_array::{Array, GenericByteArray, types::Utf8Type};
63/// let arr: GenericByteArray<Utf8Type> = (0..5).map(|x| (x % 2 == 0).then(|| x.to_string())).collect();
64/// let values: Vec<_> = arr.iter().collect();
65/// assert_eq!(values, &[Some("0"), None, Some("2"), None, Some("4")]);
66/// ```
67///
68/// # Example: Using Builder
69///
70/// ```
71/// # use arrow_array::Array;
72/// # use arrow_array::builder::GenericByteBuilder;
73/// # use arrow_array::types::Utf8Type;
74/// let mut builder = GenericByteBuilder::<Utf8Type>::new();
75/// builder.append_value("hello");
76/// builder.append_null();
77/// builder.append_value("world");
78/// let array = builder.finish();
79/// let values: Vec<_> = array.iter().collect();
80/// assert_eq!(values, &[Some("hello"), None, Some("world")]);
81/// ```
82///
83/// [`StringArray`]: crate::StringArray
84/// [`LargeStringArray`]: crate::LargeStringArray
85/// [`BinaryArray`]: crate::BinaryArray
86/// [`LargeBinaryArray`]: crate::LargeBinaryArray
87pub struct GenericByteArray<T: ByteArrayType> {
88    data_type: DataType,
89    value_offsets: OffsetBuffer<T::Offset>,
90    value_data: Buffer,
91    nulls: Option<NullBuffer>,
92}
93
94impl<T: ByteArrayType> Clone for GenericByteArray<T> {
95    fn clone(&self) -> Self {
96        Self {
97            data_type: T::DATA_TYPE,
98            value_offsets: self.value_offsets.clone(),
99            value_data: self.value_data.clone(),
100            nulls: self.nulls.clone(),
101        }
102    }
103}
104
105impl<T: ByteArrayType> GenericByteArray<T> {
106    /// Data type of the array.
107    pub const DATA_TYPE: DataType = T::DATA_TYPE;
108
109    /// Create a new [`GenericByteArray`] from the provided parts, panicking on failure
110    ///
111    /// # Panics
112    ///
113    /// Panics if [`GenericByteArray::try_new`] returns an error
114    pub fn new(
115        offsets: OffsetBuffer<T::Offset>,
116        values: Buffer,
117        nulls: Option<NullBuffer>,
118    ) -> Self {
119        Self::try_new(offsets, values, nulls).unwrap()
120    }
121
122    /// Create a new [`GenericByteArray`] from the provided parts, returning an error on failure
123    ///
124    /// # Errors
125    ///
126    /// * `offsets.len() - 1 != nulls.len()`
127    /// * Any consecutive pair of `offsets` does not denote a valid slice of `values`
128    pub fn try_new(
129        offsets: OffsetBuffer<T::Offset>,
130        values: Buffer,
131        nulls: Option<NullBuffer>,
132    ) -> Result<Self, ArrowError> {
133        let len = offsets.len() - 1;
134
135        // Verify that each pair of offsets is a valid slices of values
136        T::validate(&offsets, &values)?;
137
138        if let Some(n) = nulls.as_ref()
139            && n.len() != len
140        {
141            return Err(ArrowError::InvalidArgumentError(format!(
142                "Incorrect length of null buffer for {}{}Array, expected {len} got {}",
143                T::Offset::PREFIX,
144                T::PREFIX,
145                n.len(),
146            )));
147        }
148
149        Ok(Self {
150            data_type: T::DATA_TYPE,
151            value_offsets: offsets,
152            value_data: values,
153            nulls,
154        })
155    }
156
157    /// Create a new [`GenericByteArray`] from the provided parts, without validation
158    ///
159    /// # Safety
160    ///
161    /// Safe if [`Self::try_new`] would not error
162    pub unsafe fn new_unchecked(
163        offsets: OffsetBuffer<T::Offset>,
164        values: Buffer,
165        nulls: Option<NullBuffer>,
166    ) -> Self {
167        if cfg!(feature = "force_validate") {
168            return Self::new(offsets, values, nulls);
169        }
170        Self {
171            data_type: T::DATA_TYPE,
172            value_offsets: offsets,
173            value_data: values,
174            nulls,
175        }
176    }
177
178    /// Create a new [`GenericByteArray`] of length `len` where all values are null
179    pub fn new_null(len: usize) -> Self {
180        Self {
181            data_type: T::DATA_TYPE,
182            value_offsets: OffsetBuffer::new_zeroed(len),
183            value_data: MutableBuffer::new(0).into(),
184            nulls: Some(NullBuffer::new_null(len)),
185        }
186    }
187
188    /// Create a new [`Scalar`] from `v`
189    pub fn new_scalar(value: impl AsRef<T::Native>) -> Scalar<Self> {
190        Scalar::new(Self::from_iter_values(std::iter::once(value)))
191    }
192
193    /// Create a new [`GenericByteArray`] where `value` is repeated `repeat_count` times.
194    ///
195    /// # Panics
196    /// This will panic if value's length multiplied by `repeat_count` overflows usize.
197    ///
198    pub fn new_repeated(value: impl AsRef<T::Native>, repeat_count: usize) -> Self {
199        let s: &[u8] = value.as_ref().as_ref();
200        let value_offsets = OffsetBuffer::from_repeated_length(s.len(), repeat_count);
201        let bytes: Buffer = {
202            let mut mutable_buffer = MutableBuffer::with_capacity(0);
203            mutable_buffer.repeat_slice_n_times(s, repeat_count);
204
205            mutable_buffer.into()
206        };
207
208        Self {
209            data_type: T::DATA_TYPE,
210            value_data: bytes,
211            value_offsets,
212            nulls: None,
213        }
214    }
215
216    /// Creates a [`GenericByteArray`] based on an iterator of values without nulls
217    ///
218    /// # Panics
219    /// Panics if the total length of the values exceeds `T::Offset::MAX`
220    pub fn from_iter_values<Ptr, I>(iter: I) -> Self
221    where
222        Ptr: AsRef<T::Native>,
223        I: IntoIterator<Item = Ptr>,
224    {
225        let iter = iter.into_iter();
226        // The size hint is only used to pre-allocate: an iterator is free to yield
227        // a different number of items than it reports.
228        let (lower, upper) = iter.size_hint();
229        let capacity = upper.unwrap_or(lower);
230
231        let mut offsets = MutableBuffer::new(
232            capacity
233                .saturating_add(1)
234                .saturating_mul(std::mem::size_of::<T::Offset>()),
235        );
236        offsets.push(T::Offset::usize_as(0));
237
238        let mut values = MutableBuffer::new(0);
239        for s in iter {
240            let s: &[u8] = s.as_ref().as_ref();
241            values.extend_from_slice(s);
242            offsets.push(T::Offset::usize_as(values.len()));
243        }
244
245        T::Offset::from_usize(values.len()).expect("offset overflow");
246        let offsets = Buffer::from(offsets);
247
248        // Safety: valid by construction
249        let value_offsets = unsafe { OffsetBuffer::new_unchecked(offsets.into()) };
250
251        Self {
252            data_type: T::DATA_TYPE,
253            value_data: values.into(),
254            value_offsets,
255            nulls: None,
256        }
257    }
258
259    /// Deconstruct this array into its constituent parts
260    pub fn into_parts(self) -> (OffsetBuffer<T::Offset>, Buffer, Option<NullBuffer>) {
261        (self.value_offsets, self.value_data, self.nulls)
262    }
263
264    /// Returns the length for value at index `i`.
265    /// # Panics
266    /// Panics if index `i` is out of bounds.
267    #[inline]
268    pub fn value_length(&self, i: usize) -> T::Offset {
269        let offsets = self.value_offsets();
270        offsets[i + 1] - offsets[i]
271    }
272
273    /// Returns a reference to the offsets of this array
274    ///
275    /// Unlike [`Self::value_offsets`] this returns the [`OffsetBuffer`]
276    /// allowing for zero-copy cloning
277    #[inline]
278    pub fn offsets(&self) -> &OffsetBuffer<T::Offset> {
279        &self.value_offsets
280    }
281
282    /// Returns the values of this array
283    ///
284    /// Unlike [`Self::value_data`] this returns the [`Buffer`]
285    /// allowing for zero-copy cloning
286    #[inline]
287    pub fn values(&self) -> &Buffer {
288        &self.value_data
289    }
290
291    /// Returns the raw value data
292    pub fn value_data(&self) -> &[u8] {
293        self.value_data.as_slice()
294    }
295
296    /// Returns true if all data within this array is ASCII
297    pub fn is_ascii(&self) -> bool {
298        let offsets = &self.value_offsets;
299        self.value_data()[offsets.first().as_usize()..offsets.last().as_usize()].is_ascii()
300    }
301
302    /// Returns the offset values in the offsets buffer
303    #[inline]
304    pub fn value_offsets(&self) -> &[T::Offset] {
305        &self.value_offsets
306    }
307
308    /// Returns the element at index `i`
309    ///
310    /// Note: This method does not check for nulls and the value is arbitrary
311    /// if [`is_null`](Self::is_null) returns true for the index.
312    ///
313    /// # Safety
314    /// Caller is responsible for ensuring that the index is within the bounds of the array
315    pub unsafe fn value_unchecked(&self, i: usize) -> &T::Native {
316        let end = *unsafe { self.value_offsets().get_unchecked(i + 1) };
317        let start = *unsafe { self.value_offsets().get_unchecked(i) };
318
319        // Soundness
320        // pointer alignment & location is ensured by RawPtrBox
321        // buffer bounds/offset is ensured by the value_offset invariants
322
323        // Safety of `to_isize().unwrap()`
324        // `start` and `end` are &OffsetSize, which is a generic type that implements the
325        // OffsetSizeTrait. Currently, only i32 and i64 implement OffsetSizeTrait,
326        // both of which should cleanly cast to isize on an architecture that supports
327        // 32/64-bit offsets
328        let b = unsafe {
329            std::slice::from_raw_parts(
330                self.value_data
331                    .as_ptr()
332                    .offset(start.to_isize().unwrap_unchecked()),
333                (end - start).to_usize().unwrap_unchecked(),
334            )
335        };
336
337        // SAFETY:
338        // ArrayData is valid
339        unsafe { T::Native::from_bytes_unchecked(b) }
340    }
341
342    /// Returns the element at index `i`
343    ///
344    /// Note: This method does not check for nulls and the value is arbitrary
345    /// (but still well-defined) if [`is_null`](Self::is_null) returns true for the index.
346    ///
347    /// # Panics
348    /// Panics if index `i` is out of bounds.
349    pub fn value(&self, i: usize) -> &T::Native {
350        assert!(
351            i < self.len(),
352            "Trying to access an element at index {} from a {}{}Array of length {}",
353            i,
354            T::Offset::PREFIX,
355            T::PREFIX,
356            self.len()
357        );
358        // SAFETY:
359        // Verified length above
360        unsafe { self.value_unchecked(i) }
361    }
362
363    /// constructs a new iterator
364    pub fn iter(&self) -> ArrayIter<&Self> {
365        ArrayIter::new(self)
366    }
367
368    /// Returns a zero-copy slice of this array with the indicated offset and length.
369    ///
370    /// # Panics
371    /// Panics if `offset + length > self.len()`
372    pub fn slice(&self, offset: usize, length: usize) -> Self {
373        Self {
374            data_type: T::DATA_TYPE,
375            value_offsets: self.value_offsets.slice(offset, length),
376            value_data: self.value_data.clone(),
377            nulls: self.nulls.as_ref().map(|n| n.slice(offset, length)),
378        }
379    }
380
381    /// Returns `GenericByteBuilder` of this byte array for mutating its values if the underlying
382    /// offset and data buffers are not shared by others.
383    pub fn into_builder(self) -> Result<GenericByteBuilder<T>, Self> {
384        let len = self.len();
385        let value_len = T::Offset::as_usize(self.value_offsets()[len] - self.value_offsets()[0]);
386
387        let data = self.into_data();
388        let null_bit_buffer = data.nulls().map(|b| b.inner().sliced());
389
390        let element_len = std::mem::size_of::<T::Offset>();
391        let offset_buffer = data.buffers()[0]
392            .slice_with_length(data.offset() * element_len, (len + 1) * element_len);
393
394        let element_len = std::mem::size_of::<u8>();
395        let value_buffer = data.buffers()[1]
396            .slice_with_length(data.offset() * element_len, value_len * element_len);
397
398        drop(data);
399
400        let try_mutable_null_buffer = match null_bit_buffer {
401            None => Ok(None),
402            Some(null_buffer) => {
403                // Null buffer exists, tries to make it mutable
404                null_buffer.into_mutable().map(Some)
405            }
406        };
407
408        let try_mutable_buffers = match try_mutable_null_buffer {
409            Ok(mutable_null_buffer) => {
410                // Got mutable null buffer, tries to get mutable value buffer
411                let try_mutable_offset_buffer = offset_buffer.into_mutable();
412                let try_mutable_value_buffer = value_buffer.into_mutable();
413
414                // try_mutable_offset_buffer.map(...).map_err(...) doesn't work as the compiler complains
415                // mutable_null_buffer is moved into map closure.
416                match (try_mutable_offset_buffer, try_mutable_value_buffer) {
417                    (Ok(mutable_offset_buffer), Ok(mutable_value_buffer)) => unsafe {
418                        Ok(GenericByteBuilder::<T>::new_from_buffer(
419                            mutable_offset_buffer,
420                            mutable_value_buffer,
421                            mutable_null_buffer,
422                        ))
423                    },
424                    (Ok(mutable_offset_buffer), Err(value_buffer)) => Err((
425                        mutable_offset_buffer.into(),
426                        value_buffer,
427                        mutable_null_buffer.map(|b| b.into()),
428                    )),
429                    (Err(offset_buffer), Ok(mutable_value_buffer)) => Err((
430                        offset_buffer,
431                        mutable_value_buffer.into(),
432                        mutable_null_buffer.map(|b| b.into()),
433                    )),
434                    (Err(offset_buffer), Err(value_buffer)) => Err((
435                        offset_buffer,
436                        value_buffer,
437                        mutable_null_buffer.map(|b| b.into()),
438                    )),
439                }
440            }
441            Err(mutable_null_buffer) => {
442                // Unable to get mutable null buffer
443                Err((offset_buffer, value_buffer, Some(mutable_null_buffer)))
444            }
445        };
446
447        match try_mutable_buffers {
448            Ok(builder) => Ok(builder),
449            Err((offset_buffer, value_buffer, null_bit_buffer)) => {
450                let builder = ArrayData::builder(T::DATA_TYPE)
451                    .len(len)
452                    .add_buffer(offset_buffer)
453                    .add_buffer(value_buffer)
454                    .null_bit_buffer(null_bit_buffer);
455
456                let array_data = unsafe { builder.build_unchecked() };
457                let array = GenericByteArray::<T>::from(array_data);
458
459                Err(array)
460            }
461        }
462    }
463}
464
465impl<T: ByteArrayType> std::fmt::Debug for GenericByteArray<T> {
466    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
467        write!(f, "{}{}Array\n[\n", T::Offset::PREFIX, T::PREFIX)?;
468        print_long_array(self, f, &mut |index, f| {
469            std::fmt::Debug::fmt(&self.value(index), f)
470        })?;
471        write!(f, "]")
472    }
473}
474
475/// SAFETY: Correctly implements the contract of Arrow Arrays
476unsafe impl<T: ByteArrayType> Array for GenericByteArray<T> {
477    fn as_any(&self) -> &dyn Any {
478        self
479    }
480
481    fn to_data(&self) -> ArrayData {
482        self.clone().into()
483    }
484
485    fn into_data(self) -> ArrayData {
486        self.into()
487    }
488
489    fn data_type(&self) -> &DataType {
490        &self.data_type
491    }
492
493    fn slice(&self, offset: usize, length: usize) -> ArrayRef {
494        Arc::new(self.slice(offset, length))
495    }
496
497    fn len(&self) -> usize {
498        self.value_offsets.len() - 1
499    }
500
501    fn is_empty(&self) -> bool {
502        self.value_offsets.len() <= 1
503    }
504
505    fn shrink_to_fit(&mut self) {
506        self.value_offsets.shrink_to_fit();
507        self.value_data.shrink_to_fit();
508        if let Some(nulls) = &mut self.nulls {
509            nulls.shrink_to_fit();
510        }
511    }
512
513    fn offset(&self) -> usize {
514        0
515    }
516
517    fn nulls(&self) -> Option<&NullBuffer> {
518        self.nulls.as_ref()
519    }
520
521    fn logical_null_count(&self) -> usize {
522        // More efficient that the default implementation
523        self.null_count()
524    }
525
526    fn get_buffer_memory_size(&self) -> usize {
527        let mut sum = self.value_offsets.inner().inner().capacity();
528        sum += self.value_data.capacity();
529        if let Some(x) = &self.nulls {
530            sum += x.buffer().capacity()
531        }
532        sum
533    }
534
535    fn get_array_memory_size(&self) -> usize {
536        std::mem::size_of::<Self>() + self.get_buffer_memory_size()
537    }
538
539    #[cfg(feature = "pool")]
540    fn claim(&self, pool: &dyn arrow_buffer::MemoryPool) {
541        self.value_offsets.claim(pool);
542        self.value_data.claim(pool);
543        if let Some(nulls) = &self.nulls {
544            nulls.claim(pool);
545        }
546    }
547}
548
549impl<'a, T: ByteArrayType> ArrayAccessor for &'a GenericByteArray<T> {
550    type Item = &'a T::Native;
551
552    fn value(&self, index: usize) -> Self::Item {
553        GenericByteArray::value(self, index)
554    }
555
556    unsafe fn value_unchecked(&self, index: usize) -> Self::Item {
557        unsafe { GenericByteArray::value_unchecked(self, index) }
558    }
559}
560
561impl<T: ByteArrayType> From<ArrayData> for GenericByteArray<T> {
562    fn from(data: ArrayData) -> Self {
563        let (data_type, len, nulls, offset, mut buffers, _child_data) = data.into_parts();
564        assert_eq!(
565            data_type,
566            Self::DATA_TYPE,
567            "{}{}Array expects DataType::{}",
568            T::Offset::PREFIX,
569            T::PREFIX,
570            Self::DATA_TYPE
571        );
572        assert_eq!(
573            buffers.len(),
574            2,
575            "{}{}Array data should contain 2 buffers only (offsets and values)",
576            T::Offset::PREFIX,
577            T::PREFIX,
578        );
579        // buffers are offset then value, so pop in reverse
580        let value_data = buffers.pop().expect("checked above");
581        let offset_buffer = buffers.pop().expect("checked above");
582
583        // SAFETY:
584        // ArrayData is valid, and verified type above
585        let value_offsets = unsafe { get_offsets_from_buffer(offset_buffer, offset, len) };
586        Self {
587            data_type,
588            value_offsets,
589            value_data,
590            nulls,
591        }
592    }
593}
594
595impl<T: ByteArrayType> From<GenericByteArray<T>> for ArrayData {
596    fn from(array: GenericByteArray<T>) -> Self {
597        let len = array.len();
598
599        let offsets = array.value_offsets.into_inner().into_inner();
600        let builder = ArrayDataBuilder::new(array.data_type)
601            .len(len)
602            .buffers(vec![offsets, array.value_data])
603            .nulls(array.nulls);
604
605        unsafe { builder.build_unchecked() }
606    }
607}
608
609impl<'a, T: ByteArrayType> IntoIterator for &'a GenericByteArray<T> {
610    type Item = Option<&'a T::Native>;
611    type IntoIter = ArrayIter<Self>;
612
613    fn into_iter(self) -> Self::IntoIter {
614        ArrayIter::new(self)
615    }
616}
617
618impl<'a, Ptr, T: ByteArrayType> FromIterator<&'a Option<Ptr>> for GenericByteArray<T>
619where
620    Ptr: AsRef<T::Native> + 'a,
621{
622    fn from_iter<I: IntoIterator<Item = &'a Option<Ptr>>>(iter: I) -> Self {
623        iter.into_iter()
624            .map(|o| o.as_ref().map(|p| p.as_ref()))
625            .collect()
626    }
627}
628
629impl<Ptr, T: ByteArrayType> FromIterator<Option<Ptr>> for GenericByteArray<T>
630where
631    Ptr: AsRef<T::Native>,
632{
633    fn from_iter<I: IntoIterator<Item = Option<Ptr>>>(iter: I) -> Self {
634        let iter = iter.into_iter();
635        let mut builder = GenericByteBuilder::with_capacity(iter.size_hint().0, 1024);
636        builder.extend(iter);
637        builder.finish()
638    }
639}
640
641#[cfg(test)]
642mod tests {
643    use crate::{Array, BinaryArray, StringArray};
644    use arrow_buffer::{Buffer, NullBuffer, OffsetBuffer};
645
646    /// `from_iter_values` must work with iterators that report no upper size bound,
647    /// and must not trust the size hint it does get.
648    #[test]
649    fn from_iter_values_untrusted_size_hint() {
650        // No upper bound at all:
651        let no_upper_bound = (0..20).filter(|i| i % 2 == 0).map(|i| format!("v{i}"));
652        assert_eq!(no_upper_bound.size_hint(), (0, Some(20)));
653        let array = StringArray::from_iter_values(no_upper_bound);
654        assert_eq!(array.len(), 10);
655        assert_eq!(array.value(0), "v0");
656        assert_eq!(array.value(9), "v18");
657
658        // Upper bound larger than the number of yielded values:
659        let too_large_upper_bound = (0..).map(|i| format!("v{i}")).take_while(|v| v != "v3");
660        assert_eq!(too_large_upper_bound.size_hint(), (0, None));
661        let array = StringArray::from_iter_values(too_large_upper_bound);
662        assert_eq!(array.len(), 3);
663        assert_eq!(array.value(2), "v2");
664    }
665
666    #[test]
667    fn try_new() {
668        let data = Buffer::from_slice_ref("helloworld");
669        let offsets = OffsetBuffer::new(vec![0, 5, 10].into());
670        StringArray::new(offsets.clone(), data.clone(), None);
671
672        let nulls = NullBuffer::new_null(3);
673        let err =
674            StringArray::try_new(offsets.clone(), data.clone(), Some(nulls.clone())).unwrap_err();
675        assert_eq!(
676            err.to_string(),
677            "Invalid argument error: Incorrect length of null buffer for StringArray, expected 2 got 3"
678        );
679
680        let err = BinaryArray::try_new(offsets.clone(), data.clone(), Some(nulls)).unwrap_err();
681        assert_eq!(
682            err.to_string(),
683            "Invalid argument error: Incorrect length of null buffer for BinaryArray, expected 2 got 3"
684        );
685
686        let non_utf8_data = Buffer::from_slice_ref(b"he\xFFloworld");
687        let err = StringArray::try_new(offsets.clone(), non_utf8_data.clone(), None).unwrap_err();
688        assert_eq!(
689            err.to_string(),
690            "Invalid argument error: Encountered non UTF-8 data: invalid utf-8 sequence of 1 bytes from index 2"
691        );
692
693        BinaryArray::new(offsets, non_utf8_data, None);
694
695        let offsets = OffsetBuffer::new(vec![0, 5, 11].into());
696        let err = StringArray::try_new(offsets.clone(), data.clone(), None).unwrap_err();
697        assert_eq!(
698            err.to_string(),
699            "Invalid argument error: Offset of 11 exceeds length of values 10"
700        );
701
702        let err = BinaryArray::try_new(offsets.clone(), data, None).unwrap_err();
703        assert_eq!(
704            err.to_string(),
705            "Invalid argument error: Maximum offset of 11 is larger than values of length 10"
706        );
707
708        let non_ascii_data = Buffer::from_slice_ref("heìloworld");
709        StringArray::new(offsets.clone(), non_ascii_data.clone(), None);
710        BinaryArray::new(offsets, non_ascii_data.clone(), None);
711
712        let offsets = OffsetBuffer::new(vec![0, 3, 10].into());
713        let err = StringArray::try_new(offsets.clone(), non_ascii_data.clone(), None).unwrap_err();
714        assert_eq!(
715            err.to_string(),
716            "Invalid argument error: Split UTF-8 codepoint at offset 3"
717        );
718
719        BinaryArray::new(offsets, non_ascii_data, None);
720    }
721
722    #[test]
723    fn create_repeated() {
724        let arr = BinaryArray::new_repeated(b"hello", 3);
725        assert_eq!(arr.len(), 3);
726        assert_eq!(arr.value(0), b"hello");
727        assert_eq!(arr.value(1), b"hello");
728        assert_eq!(arr.value(2), b"hello");
729
730        let arr = StringArray::new_repeated("world", 2);
731        assert_eq!(arr.len(), 2);
732        assert_eq!(arr.value(0), "world");
733        assert_eq!(arr.value(1), "world");
734    }
735
736    #[test]
737    #[should_panic(expected = "total length overflow: does not fit in usize")]
738    fn create_repeated_usize_overflow_1() {
739        let _arr = BinaryArray::new_repeated(b"hello", (usize::MAX / "hello".len()) + 1);
740    }
741
742    #[test]
743    #[should_panic(expected = "total length overflow: does not fit in usize")]
744    fn create_repeated_usize_overflow_2() {
745        let _arr = BinaryArray::new_repeated(b"hello", usize::MAX);
746    }
747
748    #[test]
749    #[should_panic(expected = "offset overflow")]
750    fn create_repeated_i32_offset_overflow_1() {
751        let _arr = BinaryArray::new_repeated(b"hello", usize::MAX / "hello".len());
752    }
753
754    #[test]
755    #[should_panic(expected = "offset overflow")]
756    fn create_repeated_i32_offset_overflow_2() {
757        let _arr = BinaryArray::new_repeated(b"hello", ((i32::MAX as usize) / "hello".len()) + 1);
758    }
759}
arrow_array/array/byte_array.rs

arrow_array/array/
byte_array.rs
