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

arrow_data/
byte_view.rs
