arrow_data/
byte_view.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 arrow_buffer::Buffer;
19use arrow_schema::ArrowError;
20
21/// The maximum number of bytes that can be stored inline in a byte view.
22///
23/// See [`ByteView`] and [`GenericByteViewArray`] for more information on the
24/// layout of the views.
25///
26/// [`GenericByteViewArray`]: https://docs.rs/arrow/latest/arrow/array/struct.GenericByteViewArray.html
27pub const MAX_INLINE_VIEW_LEN: u32 = 12;
28
29/// Helper to access views of [`GenericByteViewArray`] (`StringViewArray` and
30/// `BinaryViewArray`) where the length is greater than 12 bytes.
31///
32/// See Also:
33/// * [`GenericByteViewArray`] for more information on the layout of the views.
34/// * [`validate_binary_view`] and [`validate_string_view`] to validate
35///
36/// # Example: Create a new u128 view
37///
38/// ```rust
39/// # use arrow_data::ByteView;;
40/// // Create a view for a string of length 20
41/// // first four bytes are "Rust"
42/// // stored in buffer 3
43/// // at offset 42
44/// let prefix = "Rust";
45/// let view = ByteView::new(20, prefix.as_bytes())
46///   .with_buffer_index(3)
47///   .with_offset(42);
48///
49/// // create the final u128
50/// let v = view.as_u128();
51/// assert_eq!(v, 0x2a000000037473755200000014);
52/// ```
53///
54/// # Example: decode a `u128` into its constituent fields
55/// ```rust
56/// # use arrow_data::ByteView;
57/// // Convert a u128 to a ByteView
58/// // See validate_{string,binary}_view functions to validate
59/// let v = ByteView::from(0x2a000000037473755200000014);
60///
61/// assert_eq!(v.length, 20);
62/// assert_eq!(v.prefix, 0x74737552);
63/// assert_eq!(v.buffer_index, 3);
64/// assert_eq!(v.offset, 42);
65/// ```
66///
67/// [`GenericByteViewArray`]: https://docs.rs/arrow/latest/arrow/array/struct.GenericByteViewArray.html
68#[derive(Debug, Copy, Clone, Default)]
69#[repr(C)]
70pub struct ByteView {
71    /// The length of the string/bytes.
72    pub length: u32,
73    /// First 4 bytes of string/bytes data.
74    pub prefix: u32,
75    /// The buffer index.
76    pub buffer_index: u32,
77    /// The offset into the buffer.
78    pub offset: u32,
79}
80
81impl ByteView {
82    /// Construct a [`ByteView`] for data `length` of bytes with the specified prefix.
83    ///
84    /// See example on [`ByteView`] docs
85    ///
86    /// Notes:
87    /// * the length should always be greater than [`MAX_INLINE_VIEW_LEN`]
88    ///   (Data less than 12 bytes is stored as an inline view)
89    /// * buffer and offset are set to `0`
90    ///
91    /// # Panics
92    /// If the prefix is not exactly 4 bytes
93    #[inline]
94    pub fn new(length: u32, prefix: &[u8]) -> Self {
95        debug_assert!(length > MAX_INLINE_VIEW_LEN);
96        Self {
97            length,
98            prefix: u32::from_le_bytes(prefix.try_into().unwrap()),
99            buffer_index: 0,
100            offset: 0,
101        }
102    }
103
104    /// Set the [`Self::buffer_index`] field
105    #[inline]
106    pub fn with_buffer_index(mut self, buffer_index: u32) -> Self {
107        self.buffer_index = buffer_index;
108        self
109    }
110
111    /// Set the [`Self::offset`] field
112    #[inline]
113    pub fn with_offset(mut self, offset: u32) -> Self {
114        self.offset = offset;
115        self
116    }
117
118    #[inline(always)]
119    /// Convert `ByteView` to `u128` by concatenating the fields
120    pub fn as_u128(self) -> u128 {
121        (self.length as u128)
122            | ((self.prefix as u128) << 32)
123            | ((self.buffer_index as u128) << 64)
124            | ((self.offset as u128) << 96)
125    }
126}
127
128impl From<u128> for ByteView {
129    #[inline]
130    fn from(value: u128) -> Self {
131        Self {
132            length: value as u32,
133            prefix: (value >> 32) as u32,
134            buffer_index: (value >> 64) as u32,
135            offset: (value >> 96) as u32,
136        }
137    }
138}
139
140impl From<ByteView> for u128 {
141    #[inline]
142    fn from(value: ByteView) -> Self {
143        value.as_u128()
144    }
145}
146
147/// Validates the combination of `views` and `buffers` is a valid BinaryView
148pub fn validate_binary_view(views: &[u128], buffers: &[Buffer]) -> Result<(), ArrowError> {
149    validate_view_impl(views, buffers, |_, _| Ok(()))
150}
151
152/// Validates the combination of `views` and `buffers` is a valid StringView
153pub fn validate_string_view(views: &[u128], buffers: &[Buffer]) -> Result<(), ArrowError> {
154    validate_view_impl(views, buffers, |idx, b| {
155        std::str::from_utf8(b).map_err(|e| {
156            ArrowError::InvalidArgumentError(format!(
157                "Encountered non-UTF-8 data at index {idx}: {e}"
158            ))
159        })?;
160        Ok(())
161    })
162}
163
164fn validate_view_impl<F>(views: &[u128], buffers: &[Buffer], f: F) -> Result<(), ArrowError>
165where
166    F: Fn(usize, &[u8]) -> Result<(), ArrowError>,
167{
168    for (idx, v) in views.iter().enumerate() {
169        let len = *v as u32;
170        if len <= MAX_INLINE_VIEW_LEN {
171            if len < MAX_INLINE_VIEW_LEN && (v >> (32 + len * 8)) != 0 {
172                return Err(ArrowError::InvalidArgumentError(format!(
173                    "View at index {idx} contained non-zero padding for string of length {len}",
174                )));
175            }
176            f(idx, &v.to_le_bytes()[4..4 + len as usize])?;
177        } else {
178            let view = ByteView::from(*v);
179            let data = buffers.get(view.buffer_index as usize).ok_or_else(|| {
180                ArrowError::InvalidArgumentError(format!(
181                    "Invalid buffer index at {idx}: got index {} but only has {} buffers",
182                    view.buffer_index,
183                    buffers.len()
184                ))
185            })?;
186
187            let start = view.offset as usize;
188            let end = start + len as usize;
189            let b = data.get(start..end).ok_or_else(|| {
190                ArrowError::InvalidArgumentError(format!(
191                    "Invalid buffer slice at {idx}: got {start}..{end} but buffer {} has length {}",
192                    view.buffer_index,
193                    data.len()
194                ))
195            })?;
196
197            if !b.starts_with(&view.prefix.to_le_bytes()) {
198                return Err(ArrowError::InvalidArgumentError(
199                    "Mismatch between embedded prefix and data".to_string(),
200                ));
201            }
202
203            f(idx, b)?;
204        }
205    }
206    Ok(())
207}
arrow_data/byte_view.rs

arrow_data/
byte_view.rs
