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 iterator has no upper bound on its size hint, or if the total
220    /// length of the values exceeds `T::Offset::MAX`
221    pub fn from_iter_values<Ptr, I>(iter: I) -> Self
222    where
223        Ptr: AsRef<T::Native>,
224        I: IntoIterator<Item = Ptr>,
225    {
226        let iter = iter.into_iter();
227        let (_, data_len) = iter.size_hint();
228        let data_len = data_len.expect("Iterator must be sized"); // panic if no upper bound.
229
230        let mut offsets = MutableBuffer::new((data_len + 1) * std::mem::size_of::<T::Offset>());
231        offsets.push(T::Offset::usize_as(0));
232
233        let mut values = MutableBuffer::new(0);
234        for s in iter {
235            let s: &[u8] = s.as_ref().as_ref();
236            values.extend_from_slice(s);
237            offsets.push(T::Offset::usize_as(values.len()));
238        }
239
240        T::Offset::from_usize(values.len()).expect("offset overflow");
241        let offsets = Buffer::from(offsets);
242
243        // Safety: valid by construction
244        let value_offsets = unsafe { OffsetBuffer::new_unchecked(offsets.into()) };
245
246        Self {
247            data_type: T::DATA_TYPE,
248            value_data: values.into(),
249            value_offsets,
250            nulls: None,
251        }
252    }
253
254    /// Deconstruct this array into its constituent parts
255    pub fn into_parts(self) -> (OffsetBuffer<T::Offset>, Buffer, Option<NullBuffer>) {
256        (self.value_offsets, self.value_data, self.nulls)
257    }
258
259    /// Returns the length for value at index `i`.
260    /// # Panics
261    /// Panics if index `i` is out of bounds.
262    #[inline]
263    pub fn value_length(&self, i: usize) -> T::Offset {
264        let offsets = self.value_offsets();
265        offsets[i + 1] - offsets[i]
266    }
267
268    /// Returns a reference to the offsets of this array
269    ///
270    /// Unlike [`Self::value_offsets`] this returns the [`OffsetBuffer`]
271    /// allowing for zero-copy cloning
272    #[inline]
273    pub fn offsets(&self) -> &OffsetBuffer<T::Offset> {
274        &self.value_offsets
275    }
276
277    /// Returns the values of this array
278    ///
279    /// Unlike [`Self::value_data`] this returns the [`Buffer`]
280    /// allowing for zero-copy cloning
281    #[inline]
282    pub fn values(&self) -> &Buffer {
283        &self.value_data
284    }
285
286    /// Returns the raw value data
287    pub fn value_data(&self) -> &[u8] {
288        self.value_data.as_slice()
289    }
290
291    /// Returns true if all data within this array is ASCII
292    pub fn is_ascii(&self) -> bool {
293        let offsets = self.value_offsets();
294        let start = offsets.first().unwrap();
295        let end = offsets.last().unwrap();
296        self.value_data()[start.as_usize()..end.as_usize()].is_ascii()
297    }
298
299    /// Returns the offset values in the offsets buffer
300    #[inline]
301    pub fn value_offsets(&self) -> &[T::Offset] {
302        &self.value_offsets
303    }
304
305    /// Returns the element at index `i`
306    ///
307    /// Note: This method does not check for nulls and the value is arbitrary
308    /// if [`is_null`](Self::is_null) returns true for the index.
309    ///
310    /// # Safety
311    /// Caller is responsible for ensuring that the index is within the bounds of the array
312    pub unsafe fn value_unchecked(&self, i: usize) -> &T::Native {
313        let end = *unsafe { self.value_offsets().get_unchecked(i + 1) };
314        let start = *unsafe { self.value_offsets().get_unchecked(i) };
315
316        // Soundness
317        // pointer alignment & location is ensured by RawPtrBox
318        // buffer bounds/offset is ensured by the value_offset invariants
319
320        // Safety of `to_isize().unwrap()`
321        // `start` and `end` are &OffsetSize, which is a generic type that implements the
322        // OffsetSizeTrait. Currently, only i32 and i64 implement OffsetSizeTrait,
323        // both of which should cleanly cast to isize on an architecture that supports
324        // 32/64-bit offsets
325        let b = unsafe {
326            std::slice::from_raw_parts(
327                self.value_data
328                    .as_ptr()
329                    .offset(start.to_isize().unwrap_unchecked()),
330                (end - start).to_usize().unwrap_unchecked(),
331            )
332        };
333
334        // SAFETY:
335        // ArrayData is valid
336        unsafe { T::Native::from_bytes_unchecked(b) }
337    }
338
339    /// Returns the element at index `i`
340    ///
341    /// Note: This method does not check for nulls and the value is arbitrary
342    /// (but still well-defined) if [`is_null`](Self::is_null) returns true for the index.
343    ///
344    /// # Panics
345    /// Panics if index `i` is out of bounds.
346    pub fn value(&self, i: usize) -> &T::Native {
347        assert!(
348            i < self.len(),
349            "Trying to access an element at index {} from a {}{}Array of length {}",
350            i,
351            T::Offset::PREFIX,
352            T::PREFIX,
353            self.len()
354        );
355        // SAFETY:
356        // Verified length above
357        unsafe { self.value_unchecked(i) }
358    }
359
360    /// constructs a new iterator
361    pub fn iter(&self) -> ArrayIter<&Self> {
362        ArrayIter::new(self)
363    }
364
365    /// Returns a zero-copy slice of this array with the indicated offset and length.
366    ///
367    /// # Panics
368    /// Panics if `offset + length > self.len()`
369    pub fn slice(&self, offset: usize, length: usize) -> Self {
370        Self {
371            data_type: T::DATA_TYPE,
372            value_offsets: self.value_offsets.slice(offset, length),
373            value_data: self.value_data.clone(),
374            nulls: self.nulls.as_ref().map(|n| n.slice(offset, length)),
375        }
376    }
377
378    /// Returns `GenericByteBuilder` of this byte array for mutating its values if the underlying
379    /// offset and data buffers are not shared by others.
380    pub fn into_builder(self) -> Result<GenericByteBuilder<T>, Self> {
381        let len = self.len();
382        let value_len = T::Offset::as_usize(self.value_offsets()[len] - self.value_offsets()[0]);
383
384        let data = self.into_data();
385        let null_bit_buffer = data.nulls().map(|b| b.inner().sliced());
386
387        let element_len = std::mem::size_of::<T::Offset>();
388        let offset_buffer = data.buffers()[0]
389            .slice_with_length(data.offset() * element_len, (len + 1) * element_len);
390
391        let element_len = std::mem::size_of::<u8>();
392        let value_buffer = data.buffers()[1]
393            .slice_with_length(data.offset() * element_len, value_len * element_len);
394
395        drop(data);
396
397        let try_mutable_null_buffer = match null_bit_buffer {
398            None => Ok(None),
399            Some(null_buffer) => {
400                // Null buffer exists, tries to make it mutable
401                null_buffer.into_mutable().map(Some)
402            }
403        };
404
405        let try_mutable_buffers = match try_mutable_null_buffer {
406            Ok(mutable_null_buffer) => {
407                // Got mutable null buffer, tries to get mutable value buffer
408                let try_mutable_offset_buffer = offset_buffer.into_mutable();
409                let try_mutable_value_buffer = value_buffer.into_mutable();
410
411                // try_mutable_offset_buffer.map(...).map_err(...) doesn't work as the compiler complains
412                // mutable_null_buffer is moved into map closure.
413                match (try_mutable_offset_buffer, try_mutable_value_buffer) {
414                    (Ok(mutable_offset_buffer), Ok(mutable_value_buffer)) => unsafe {
415                        Ok(GenericByteBuilder::<T>::new_from_buffer(
416                            mutable_offset_buffer,
417                            mutable_value_buffer,
418                            mutable_null_buffer,
419                        ))
420                    },
421                    (Ok(mutable_offset_buffer), Err(value_buffer)) => Err((
422                        mutable_offset_buffer.into(),
423                        value_buffer,
424                        mutable_null_buffer.map(|b| b.into()),
425                    )),
426                    (Err(offset_buffer), Ok(mutable_value_buffer)) => Err((
427                        offset_buffer,
428                        mutable_value_buffer.into(),
429                        mutable_null_buffer.map(|b| b.into()),
430                    )),
431                    (Err(offset_buffer), Err(value_buffer)) => Err((
432                        offset_buffer,
433                        value_buffer,
434                        mutable_null_buffer.map(|b| b.into()),
435                    )),
436                }
437            }
438            Err(mutable_null_buffer) => {
439                // Unable to get mutable null buffer
440                Err((offset_buffer, value_buffer, Some(mutable_null_buffer)))
441            }
442        };
443
444        match try_mutable_buffers {
445            Ok(builder) => Ok(builder),
446            Err((offset_buffer, value_buffer, null_bit_buffer)) => {
447                let builder = ArrayData::builder(T::DATA_TYPE)
448                    .len(len)
449                    .add_buffer(offset_buffer)
450                    .add_buffer(value_buffer)
451                    .null_bit_buffer(null_bit_buffer);
452
453                let array_data = unsafe { builder.build_unchecked() };
454                let array = GenericByteArray::<T>::from(array_data);
455
456                Err(array)
457            }
458        }
459    }
460}
461
462impl<T: ByteArrayType> std::fmt::Debug for GenericByteArray<T> {
463    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
464        write!(f, "{}{}Array\n[\n", T::Offset::PREFIX, T::PREFIX)?;
465        print_long_array(self, f, |array, index, f| {
466            std::fmt::Debug::fmt(&array.value(index), f)
467        })?;
468        write!(f, "]")
469    }
470}
471
472/// SAFETY: Correctly implements the contract of Arrow Arrays
473unsafe impl<T: ByteArrayType> Array for GenericByteArray<T> {
474    fn as_any(&self) -> &dyn Any {
475        self
476    }
477
478    fn to_data(&self) -> ArrayData {
479        self.clone().into()
480    }
481
482    fn into_data(self) -> ArrayData {
483        self.into()
484    }
485
486    fn data_type(&self) -> &DataType {
487        &self.data_type
488    }
489
490    fn slice(&self, offset: usize, length: usize) -> ArrayRef {
491        Arc::new(self.slice(offset, length))
492    }
493
494    fn len(&self) -> usize {
495        self.value_offsets.len() - 1
496    }
497
498    fn is_empty(&self) -> bool {
499        self.value_offsets.len() <= 1
500    }
501
502    fn shrink_to_fit(&mut self) {
503        self.value_offsets.shrink_to_fit();
504        self.value_data.shrink_to_fit();
505        if let Some(nulls) = &mut self.nulls {
506            nulls.shrink_to_fit();
507        }
508    }
509
510    fn offset(&self) -> usize {
511        0
512    }
513
514    fn nulls(&self) -> Option<&NullBuffer> {
515        self.nulls.as_ref()
516    }
517
518    fn logical_null_count(&self) -> usize {
519        // More efficient that the default implementation
520        self.null_count()
521    }
522
523    fn get_buffer_memory_size(&self) -> usize {
524        let mut sum = self.value_offsets.inner().inner().capacity();
525        sum += self.value_data.capacity();
526        if let Some(x) = &self.nulls {
527            sum += x.buffer().capacity()
528        }
529        sum
530    }
531
532    fn get_array_memory_size(&self) -> usize {
533        std::mem::size_of::<Self>() + self.get_buffer_memory_size()
534    }
535
536    #[cfg(feature = "pool")]
537    fn claim(&self, pool: &dyn arrow_buffer::MemoryPool) {
538        self.value_offsets.claim(pool);
539        self.value_data.claim(pool);
540        if let Some(nulls) = &self.nulls {
541            nulls.claim(pool);
542        }
543    }
544}
545
546impl<'a, T: ByteArrayType> ArrayAccessor for &'a GenericByteArray<T> {
547    type Item = &'a T::Native;
548
549    fn value(&self, index: usize) -> Self::Item {
550        GenericByteArray::value(self, index)
551    }
552
553    unsafe fn value_unchecked(&self, index: usize) -> Self::Item {
554        unsafe { GenericByteArray::value_unchecked(self, index) }
555    }
556}
557
558impl<T: ByteArrayType> From<ArrayData> for GenericByteArray<T> {
559    fn from(data: ArrayData) -> Self {
560        let (data_type, len, nulls, offset, mut buffers, _child_data) = data.into_parts();
561        assert_eq!(
562            data_type,
563            Self::DATA_TYPE,
564            "{}{}Array expects DataType::{}",
565            T::Offset::PREFIX,
566            T::PREFIX,
567            Self::DATA_TYPE
568        );
569        assert_eq!(
570            buffers.len(),
571            2,
572            "{}{}Array data should contain 2 buffers only (offsets and values)",
573            T::Offset::PREFIX,
574            T::PREFIX,
575        );
576        // buffers are offset then value, so pop in reverse
577        let value_data = buffers.pop().expect("checked above");
578        let offset_buffer = buffers.pop().expect("checked above");
579
580        // SAFETY:
581        // ArrayData is valid, and verified type above
582        let value_offsets = unsafe { get_offsets_from_buffer(offset_buffer, offset, len) };
583        Self {
584            data_type,
585            value_offsets,
586            value_data,
587            nulls,
588        }
589    }
590}
591
592impl<T: ByteArrayType> From<GenericByteArray<T>> for ArrayData {
593    fn from(array: GenericByteArray<T>) -> Self {
594        let len = array.len();
595
596        let offsets = array.value_offsets.into_inner().into_inner();
597        let builder = ArrayDataBuilder::new(array.data_type)
598            .len(len)
599            .buffers(vec![offsets, array.value_data])
600            .nulls(array.nulls);
601
602        unsafe { builder.build_unchecked() }
603    }
604}
605
606impl<'a, T: ByteArrayType> IntoIterator for &'a GenericByteArray<T> {
607    type Item = Option<&'a T::Native>;
608    type IntoIter = ArrayIter<Self>;
609
610    fn into_iter(self) -> Self::IntoIter {
611        ArrayIter::new(self)
612    }
613}
614
615impl<'a, Ptr, T: ByteArrayType> FromIterator<&'a Option<Ptr>> for GenericByteArray<T>
616where
617    Ptr: AsRef<T::Native> + 'a,
618{
619    fn from_iter<I: IntoIterator<Item = &'a Option<Ptr>>>(iter: I) -> Self {
620        iter.into_iter()
621            .map(|o| o.as_ref().map(|p| p.as_ref()))
622            .collect()
623    }
624}
625
626impl<Ptr, T: ByteArrayType> FromIterator<Option<Ptr>> for GenericByteArray<T>
627where
628    Ptr: AsRef<T::Native>,
629{
630    fn from_iter<I: IntoIterator<Item = Option<Ptr>>>(iter: I) -> Self {
631        let iter = iter.into_iter();
632        let mut builder = GenericByteBuilder::with_capacity(iter.size_hint().0, 1024);
633        builder.extend(iter);
634        builder.finish()
635    }
636}
637
638#[cfg(test)]
639mod tests {
640    use crate::{Array, BinaryArray, StringArray};
641    use arrow_buffer::{Buffer, NullBuffer, OffsetBuffer};
642
643    #[test]
644    fn try_new() {
645        let data = Buffer::from_slice_ref("helloworld");
646        let offsets = OffsetBuffer::new(vec![0, 5, 10].into());
647        StringArray::new(offsets.clone(), data.clone(), None);
648
649        let nulls = NullBuffer::new_null(3);
650        let err =
651            StringArray::try_new(offsets.clone(), data.clone(), Some(nulls.clone())).unwrap_err();
652        assert_eq!(
653            err.to_string(),
654            "Invalid argument error: Incorrect length of null buffer for StringArray, expected 2 got 3"
655        );
656
657        let err = BinaryArray::try_new(offsets.clone(), data.clone(), Some(nulls)).unwrap_err();
658        assert_eq!(
659            err.to_string(),
660            "Invalid argument error: Incorrect length of null buffer for BinaryArray, expected 2 got 3"
661        );
662
663        let non_utf8_data = Buffer::from_slice_ref(b"he\xFFloworld");
664        let err = StringArray::try_new(offsets.clone(), non_utf8_data.clone(), None).unwrap_err();
665        assert_eq!(
666            err.to_string(),
667            "Invalid argument error: Encountered non UTF-8 data: invalid utf-8 sequence of 1 bytes from index 2"
668        );
669
670        BinaryArray::new(offsets, non_utf8_data, None);
671
672        let offsets = OffsetBuffer::new(vec![0, 5, 11].into());
673        let err = StringArray::try_new(offsets.clone(), data.clone(), None).unwrap_err();
674        assert_eq!(
675            err.to_string(),
676            "Invalid argument error: Offset of 11 exceeds length of values 10"
677        );
678
679        let err = BinaryArray::try_new(offsets.clone(), data, None).unwrap_err();
680        assert_eq!(
681            err.to_string(),
682            "Invalid argument error: Maximum offset of 11 is larger than values of length 10"
683        );
684
685        let non_ascii_data = Buffer::from_slice_ref("heìloworld");
686        StringArray::new(offsets.clone(), non_ascii_data.clone(), None);
687        BinaryArray::new(offsets, non_ascii_data.clone(), None);
688
689        let offsets = OffsetBuffer::new(vec![0, 3, 10].into());
690        let err = StringArray::try_new(offsets.clone(), non_ascii_data.clone(), None).unwrap_err();
691        assert_eq!(
692            err.to_string(),
693            "Invalid argument error: Split UTF-8 codepoint at offset 3"
694        );
695
696        BinaryArray::new(offsets, non_ascii_data, None);
697    }
698
699    #[test]
700    fn create_repeated() {
701        let arr = BinaryArray::new_repeated(b"hello", 3);
702        assert_eq!(arr.len(), 3);
703        assert_eq!(arr.value(0), b"hello");
704        assert_eq!(arr.value(1), b"hello");
705        assert_eq!(arr.value(2), b"hello");
706
707        let arr = StringArray::new_repeated("world", 2);
708        assert_eq!(arr.len(), 2);
709        assert_eq!(arr.value(0), "world");
710        assert_eq!(arr.value(1), "world");
711    }
712
713    #[test]
714    #[should_panic(expected = "usize overflow")]
715    fn create_repeated_usize_overflow_1() {
716        let _arr = BinaryArray::new_repeated(b"hello", (usize::MAX / "hello".len()) + 1);
717    }
718
719    #[test]
720    #[should_panic(expected = "usize overflow")]
721    fn create_repeated_usize_overflow_2() {
722        let _arr = BinaryArray::new_repeated(b"hello", usize::MAX);
723    }
724
725    #[test]
726    #[should_panic(expected = "offset overflow")]
727    fn create_repeated_i32_offset_overflow_1() {
728        let _arr = BinaryArray::new_repeated(b"hello", usize::MAX / "hello".len());
729    }
730
731    #[test]
732    #[should_panic(expected = "offset overflow")]
733    fn create_repeated_i32_offset_overflow_2() {
734        let _arr = BinaryArray::new_repeated(b"hello", ((i32::MAX as usize) / "hello".len()) + 1);
735    }
736}
arrow_array/array/byte_array.rs

arrow_array/array/
byte_array.rs
