1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.
//! Contains Parquet Page definitions and page reader interface.
use bytes::Bytes;
use crate::basic::{Encoding, PageType};
use crate::errors::{ParquetError, Result};
use crate::file::{metadata::ColumnChunkMetaData, statistics::Statistics};
use crate::format::PageHeader;
/// Parquet Page definition.
///
/// List of supported pages.
/// These are 1-to-1 mapped from the equivalent Thrift definitions, except `buf` which
/// used to store uncompressed bytes of the page.
#[derive(Clone)]
pub enum Page {
DataPage {
buf: Bytes,
num_values: u32,
encoding: Encoding,
def_level_encoding: Encoding,
rep_level_encoding: Encoding,
statistics: Option<Statistics>,
},
DataPageV2 {
buf: Bytes,
num_values: u32,
encoding: Encoding,
num_nulls: u32,
num_rows: u32,
def_levels_byte_len: u32,
rep_levels_byte_len: u32,
is_compressed: bool,
statistics: Option<Statistics>,
},
DictionaryPage {
buf: Bytes,
num_values: u32,
encoding: Encoding,
is_sorted: bool,
},
}
impl Page {
/// Returns [`PageType`] for this page.
pub fn page_type(&self) -> PageType {
match self {
Page::DataPage { .. } => PageType::DATA_PAGE,
Page::DataPageV2 { .. } => PageType::DATA_PAGE_V2,
Page::DictionaryPage { .. } => PageType::DICTIONARY_PAGE,
}
}
/// Returns internal byte buffer reference for this page.
pub fn buffer(&self) -> &Bytes {
match self {
Page::DataPage { ref buf, .. } => buf,
Page::DataPageV2 { ref buf, .. } => buf,
Page::DictionaryPage { ref buf, .. } => buf,
}
}
/// Returns number of values in this page.
pub fn num_values(&self) -> u32 {
match self {
Page::DataPage { num_values, .. } => *num_values,
Page::DataPageV2 { num_values, .. } => *num_values,
Page::DictionaryPage { num_values, .. } => *num_values,
}
}
/// Returns this page [`Encoding`].
pub fn encoding(&self) -> Encoding {
match self {
Page::DataPage { encoding, .. } => *encoding,
Page::DataPageV2 { encoding, .. } => *encoding,
Page::DictionaryPage { encoding, .. } => *encoding,
}
}
/// Returns optional [`Statistics`].
pub fn statistics(&self) -> Option<&Statistics> {
match self {
Page::DataPage { ref statistics, .. } => statistics.as_ref(),
Page::DataPageV2 { ref statistics, .. } => statistics.as_ref(),
Page::DictionaryPage { .. } => None,
}
}
}
/// Helper struct to represent pages with potentially compressed buffer (data page v1) or
/// compressed and concatenated buffer (def levels + rep levels + compressed values for
/// data page v2).
///
/// The difference with `Page` is that `Page` buffer is always uncompressed.
pub struct CompressedPage {
compressed_page: Page,
uncompressed_size: usize,
}
impl CompressedPage {
/// Creates `CompressedPage` from a page with potentially compressed buffer and
/// uncompressed size.
pub fn new(compressed_page: Page, uncompressed_size: usize) -> Self {
Self {
compressed_page,
uncompressed_size,
}
}
/// Returns page type.
pub fn page_type(&self) -> PageType {
self.compressed_page.page_type()
}
/// Returns underlying page with potentially compressed buffer.
pub fn compressed_page(&self) -> &Page {
&self.compressed_page
}
/// Returns uncompressed size in bytes.
pub fn uncompressed_size(&self) -> usize {
self.uncompressed_size
}
/// Returns compressed size in bytes.
///
/// Note that it is assumed that buffer is compressed, but it may not be. In this
/// case compressed size will be equal to uncompressed size.
pub fn compressed_size(&self) -> usize {
self.compressed_page.buffer().len()
}
/// Number of values in page.
pub fn num_values(&self) -> u32 {
self.compressed_page.num_values()
}
/// Returns encoding for values in page.
pub fn encoding(&self) -> Encoding {
self.compressed_page.encoding()
}
/// Returns slice of compressed buffer in the page.
pub fn data(&self) -> &[u8] {
self.compressed_page.buffer()
}
/// Returns the thrift page header
pub(crate) fn to_thrift_header(&self) -> PageHeader {
let uncompressed_size = self.uncompressed_size();
let compressed_size = self.compressed_size();
let num_values = self.num_values();
let encoding = self.encoding();
let page_type = self.page_type();
let mut page_header = PageHeader {
type_: page_type.into(),
uncompressed_page_size: uncompressed_size as i32,
compressed_page_size: compressed_size as i32,
// TODO: Add support for crc checksum
crc: None,
data_page_header: None,
index_page_header: None,
dictionary_page_header: None,
data_page_header_v2: None,
};
match self.compressed_page {
Page::DataPage {
def_level_encoding,
rep_level_encoding,
ref statistics,
..
} => {
let data_page_header = crate::format::DataPageHeader {
num_values: num_values as i32,
encoding: encoding.into(),
definition_level_encoding: def_level_encoding.into(),
repetition_level_encoding: rep_level_encoding.into(),
statistics: crate::file::statistics::to_thrift(statistics.as_ref()),
};
page_header.data_page_header = Some(data_page_header);
}
Page::DataPageV2 {
num_nulls,
num_rows,
def_levels_byte_len,
rep_levels_byte_len,
is_compressed,
ref statistics,
..
} => {
let data_page_header_v2 = crate::format::DataPageHeaderV2 {
num_values: num_values as i32,
num_nulls: num_nulls as i32,
num_rows: num_rows as i32,
encoding: encoding.into(),
definition_levels_byte_length: def_levels_byte_len as i32,
repetition_levels_byte_length: rep_levels_byte_len as i32,
is_compressed: Some(is_compressed),
statistics: crate::file::statistics::to_thrift(statistics.as_ref()),
};
page_header.data_page_header_v2 = Some(data_page_header_v2);
}
Page::DictionaryPage { is_sorted, .. } => {
let dictionary_page_header = crate::format::DictionaryPageHeader {
num_values: num_values as i32,
encoding: encoding.into(),
is_sorted: Some(is_sorted),
};
page_header.dictionary_page_header = Some(dictionary_page_header);
}
}
page_header
}
}
/// Contains page write metrics.
pub struct PageWriteSpec {
pub page_type: PageType,
pub uncompressed_size: usize,
pub compressed_size: usize,
pub num_values: u32,
pub offset: u64,
pub bytes_written: u64,
}
impl Default for PageWriteSpec {
fn default() -> Self {
Self::new()
}
}
impl PageWriteSpec {
/// Creates new spec with default page write metrics.
pub fn new() -> Self {
Self {
page_type: PageType::DATA_PAGE,
uncompressed_size: 0,
compressed_size: 0,
num_values: 0,
offset: 0,
bytes_written: 0,
}
}
}
/// Contains metadata for a page
#[derive(Clone)]
pub struct PageMetadata {
/// The number of rows within the page if known
pub num_rows: Option<usize>,
/// The number of levels within the page if known
pub num_levels: Option<usize>,
/// Returns true if the page is a dictionary page
pub is_dict: bool,
}
impl TryFrom<&PageHeader> for PageMetadata {
type Error = ParquetError;
fn try_from(value: &PageHeader) -> std::result::Result<Self, Self::Error> {
match value.type_ {
crate::format::PageType::DATA_PAGE => {
let header = value.data_page_header.as_ref().unwrap();
Ok(PageMetadata {
num_rows: None,
num_levels: Some(header.num_values as _),
is_dict: false,
})
}
crate::format::PageType::DICTIONARY_PAGE => Ok(PageMetadata {
num_rows: None,
num_levels: None,
is_dict: true,
}),
crate::format::PageType::DATA_PAGE_V2 => {
let header = value.data_page_header_v2.as_ref().unwrap();
Ok(PageMetadata {
num_rows: Some(header.num_rows as _),
num_levels: Some(header.num_values as _),
is_dict: false,
})
}
other => Err(ParquetError::General(format!(
"page type {other:?} cannot be converted to PageMetadata"
))),
}
}
}
/// API for reading pages from a column chunk.
/// This offers a iterator like API to get the next page.
pub trait PageReader: Iterator<Item = Result<Page>> + Send {
/// Gets the next page in the column chunk associated with this reader.
/// Returns `None` if there are no pages left.
fn get_next_page(&mut self) -> Result<Option<Page>>;
/// Gets metadata about the next page, returns an error if no
/// column index information
fn peek_next_page(&mut self) -> Result<Option<PageMetadata>>;
/// Skips reading the next page, returns an error if no
/// column index information
fn skip_next_page(&mut self) -> Result<()>;
/// Returns `true` if the next page can be assumed to contain the start of a new record
///
/// Prior to parquet V2 the specification was ambiguous as to whether a single record
/// could be split across multiple pages, and prior to [(#4327)] the Rust writer would do
/// this in certain situations. However, correctly interpreting the offset index relies on
/// this assumption holding [(#4943)], and so this mechanism is provided for a [`PageReader`]
/// to signal this to the calling context
///
/// [(#4327)]: https://github.com/apache/arrow-rs/pull/4327
/// [(#4943)]: https://github.com/apache/arrow-rs/pull/4943
fn at_record_boundary(&mut self) -> Result<bool> {
Ok(self.peek_next_page()?.is_none())
}
}
/// API for writing pages in a column chunk.
///
/// It is reasonable to assume that all pages will be written in the correct order, e.g.
/// dictionary page followed by data pages, or a set of data pages, etc.
pub trait PageWriter: Send {
/// Writes a page into the output stream/sink.
/// Returns `PageWriteSpec` that contains information about written page metrics,
/// including number of bytes, size, number of values, offset, etc.
///
/// This method is called for every compressed page we write into underlying buffer,
/// either data page or dictionary page.
fn write_page(&mut self, page: CompressedPage) -> Result<PageWriteSpec>;
/// Writes column chunk metadata into the output stream/sink.
///
/// This method is called once before page writer is closed, normally when writes are
/// finalised in column writer.
fn write_metadata(&mut self, metadata: &ColumnChunkMetaData) -> Result<()>;
/// Closes resources and flushes underlying sink.
/// Page writer should not be used after this method is called.
fn close(&mut self) -> Result<()>;
}
/// An iterator over pages of one specific column in a parquet file.
pub trait PageIterator: Iterator<Item = Result<Box<dyn PageReader>>> + Send {}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_page() {
let data_page = Page::DataPage {
buf: Bytes::from(vec![0, 1, 2]),
num_values: 10,
encoding: Encoding::PLAIN,
def_level_encoding: Encoding::RLE,
rep_level_encoding: Encoding::RLE,
statistics: Some(Statistics::int32(Some(1), Some(2), None, 1, true)),
};
assert_eq!(data_page.page_type(), PageType::DATA_PAGE);
assert_eq!(data_page.buffer(), vec![0, 1, 2].as_slice());
assert_eq!(data_page.num_values(), 10);
assert_eq!(data_page.encoding(), Encoding::PLAIN);
assert_eq!(
data_page.statistics(),
Some(&Statistics::int32(Some(1), Some(2), None, 1, true))
);
let data_page_v2 = Page::DataPageV2 {
buf: Bytes::from(vec![0, 1, 2]),
num_values: 10,
encoding: Encoding::PLAIN,
num_nulls: 5,
num_rows: 20,
def_levels_byte_len: 30,
rep_levels_byte_len: 40,
is_compressed: false,
statistics: Some(Statistics::int32(Some(1), Some(2), None, 1, true)),
};
assert_eq!(data_page_v2.page_type(), PageType::DATA_PAGE_V2);
assert_eq!(data_page_v2.buffer(), vec![0, 1, 2].as_slice());
assert_eq!(data_page_v2.num_values(), 10);
assert_eq!(data_page_v2.encoding(), Encoding::PLAIN);
assert_eq!(
data_page_v2.statistics(),
Some(&Statistics::int32(Some(1), Some(2), None, 1, true))
);
let dict_page = Page::DictionaryPage {
buf: Bytes::from(vec![0, 1, 2]),
num_values: 10,
encoding: Encoding::PLAIN,
is_sorted: false,
};
assert_eq!(dict_page.page_type(), PageType::DICTIONARY_PAGE);
assert_eq!(dict_page.buffer(), vec![0, 1, 2].as_slice());
assert_eq!(dict_page.num_values(), 10);
assert_eq!(dict_page.encoding(), Encoding::PLAIN);
assert_eq!(dict_page.statistics(), None);
}
#[test]
fn test_compressed_page() {
let data_page = Page::DataPage {
buf: Bytes::from(vec![0, 1, 2]),
num_values: 10,
encoding: Encoding::PLAIN,
def_level_encoding: Encoding::RLE,
rep_level_encoding: Encoding::RLE,
statistics: Some(Statistics::int32(Some(1), Some(2), None, 1, true)),
};
let cpage = CompressedPage::new(data_page, 5);
assert_eq!(cpage.page_type(), PageType::DATA_PAGE);
assert_eq!(cpage.uncompressed_size(), 5);
assert_eq!(cpage.compressed_size(), 3);
assert_eq!(cpage.num_values(), 10);
assert_eq!(cpage.encoding(), Encoding::PLAIN);
assert_eq!(cpage.data(), &[0, 1, 2]);
}
}