parquet/file/metadata/
memory.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
18//! Memory calculations for [`ParquetMetadata::memory_size`]
19//!
20//! [`ParquetMetadata::memory_size`]: crate::file::metadata::ParquetMetaData::memory_size
21use crate::basic::{BoundaryOrder, ColumnOrder, Compression, Encoding, PageType};
22use crate::data_type::private::ParquetValueType;
23use crate::file::metadata::{
24    ColumnChunkMetaData, FileMetaData, KeyValue, PageEncodingStats, ParquetPageEncodingStats,
25    RowGroupMetaData, SortingColumn,
26};
27use crate::file::page_index::column_index::{
28    ByteArrayColumnIndex, ColumnIndex, ColumnIndexMetaData, PrimitiveColumnIndex,
29};
30use crate::file::page_index::offset_index::{OffsetIndexMetaData, PageLocation};
31use crate::file::statistics::{Statistics, ValueStatistics};
32use std::collections::HashMap;
33use std::sync::Arc;
34
35/// Trait for calculating the size of various containers
36pub trait HeapSize {
37    /// Return the size of any bytes allocated on the heap by this object,
38    /// including heap memory in those structures
39    ///
40    /// Note that the size of the type itself is not included in the result --
41    /// instead, that size is added by the caller (e.g. container).
42    fn heap_size(&self) -> usize;
43}
44
45impl<T: HeapSize> HeapSize for Vec<T> {
46    fn heap_size(&self) -> usize {
47        let item_size = std::mem::size_of::<T>();
48        // account for the contents of the Vec
49        (self.capacity() * item_size) +
50        // add any heap allocations by contents
51        self.iter().map(|t| t.heap_size()).sum::<usize>()
52    }
53}
54
55impl<K: HeapSize, V: HeapSize> HeapSize for HashMap<K, V> {
56    fn heap_size(&self) -> usize {
57        let capacity = self.capacity();
58        if capacity == 0 {
59            return 0;
60        }
61
62        // HashMap doesn't provide a way to get its heap size, so this is an approximation based on
63        // the behavior of hashbrown::HashMap as at version 0.16.0, and may become inaccurate
64        // if the implementation changes.
65        let key_val_size = std::mem::size_of::<(K, V)>();
66        // Overhead for the control tags group, which may be smaller depending on architecture
67        let group_size = 16;
68        // 1 byte of metadata stored per bucket.
69        let metadata_size = 1;
70
71        // Compute the number of buckets for the capacity. Based on hashbrown's capacity_to_buckets
72        let buckets = if capacity < 15 {
73            let min_cap = match key_val_size {
74                0..=1 => 14,
75                2..=3 => 7,
76                _ => 3,
77            };
78            let cap = min_cap.max(capacity);
79            if cap < 4 {
80                4
81            } else if cap < 8 {
82                8
83            } else {
84                16
85            }
86        } else {
87            (capacity.saturating_mul(8) / 7).next_power_of_two()
88        };
89
90        group_size
91            + (buckets * (key_val_size + metadata_size))
92            + self.keys().map(|k| k.heap_size()).sum::<usize>()
93            + self.values().map(|v| v.heap_size()).sum::<usize>()
94    }
95}
96
97impl<T: HeapSize> HeapSize for Arc<T> {
98    fn heap_size(&self) -> usize {
99        // Arc stores weak and strong counts on the heap alongside an instance of T
100        2 * std::mem::size_of::<usize>() + std::mem::size_of::<T>() + self.as_ref().heap_size()
101    }
102}
103
104impl HeapSize for Arc<dyn HeapSize> {
105    fn heap_size(&self) -> usize {
106        2 * std::mem::size_of::<usize>()
107            + std::mem::size_of_val(self.as_ref())
108            + self.as_ref().heap_size()
109    }
110}
111
112impl<T: HeapSize> HeapSize for Box<T> {
113    fn heap_size(&self) -> usize {
114        std::mem::size_of::<T>() + self.as_ref().heap_size()
115    }
116}
117
118impl<T: HeapSize> HeapSize for Option<T> {
119    fn heap_size(&self) -> usize {
120        self.as_ref().map(|inner| inner.heap_size()).unwrap_or(0)
121    }
122}
123
124impl HeapSize for String {
125    fn heap_size(&self) -> usize {
126        self.capacity()
127    }
128}
129
130impl HeapSize for FileMetaData {
131    fn heap_size(&self) -> usize {
132        #[cfg(feature = "encryption")]
133        let encryption_heap_size =
134            self.encryption_algorithm.heap_size() + self.footer_signing_key_metadata.heap_size();
135        #[cfg(not(feature = "encryption"))]
136        let encryption_heap_size = 0;
137
138        self.created_by.heap_size()
139            + self.key_value_metadata.heap_size()
140            + self.schema_descr.heap_size()
141            + self.column_orders.heap_size()
142            + encryption_heap_size
143    }
144}
145
146impl HeapSize for KeyValue {
147    fn heap_size(&self) -> usize {
148        self.key.heap_size() + self.value.heap_size()
149    }
150}
151
152impl HeapSize for RowGroupMetaData {
153    fn heap_size(&self) -> usize {
154        // don't count schema_descr here because it is already
155        // counted in FileMetaData
156        self.columns.heap_size() + self.sorting_columns.heap_size()
157    }
158}
159
160impl HeapSize for ColumnChunkMetaData {
161    fn heap_size(&self) -> usize {
162        #[cfg(feature = "encryption")]
163        let encryption_heap_size =
164            self.column_crypto_metadata.heap_size() + self.encrypted_column_metadata.heap_size();
165        #[cfg(not(feature = "encryption"))]
166        let encryption_heap_size = 0;
167
168        // don't count column_descr here because it is already counted in
169        // FileMetaData
170        self.encodings.heap_size()
171            + self.file_path.heap_size()
172            + self.compression.heap_size()
173            + self.statistics.heap_size()
174            + self.encoding_stats.heap_size()
175            + self.unencoded_byte_array_data_bytes.heap_size()
176            + self.repetition_level_histogram.heap_size()
177            + self.definition_level_histogram.heap_size()
178            + self.geo_statistics.heap_size()
179            + encryption_heap_size
180    }
181}
182
183impl HeapSize for Encoding {
184    fn heap_size(&self) -> usize {
185        0 // no heap allocations
186    }
187}
188
189impl HeapSize for ParquetPageEncodingStats {
190    fn heap_size(&self) -> usize {
191        match self {
192            Self::Full(v) => v.heap_size(),
193            Self::Mask(_) => 0,
194        }
195    }
196}
197
198impl HeapSize for PageEncodingStats {
199    fn heap_size(&self) -> usize {
200        self.page_type.heap_size() + self.encoding.heap_size()
201    }
202}
203
204impl HeapSize for SortingColumn {
205    fn heap_size(&self) -> usize {
206        0 // no heap allocations
207    }
208}
209impl HeapSize for Compression {
210    fn heap_size(&self) -> usize {
211        0 // no heap allocations
212    }
213}
214
215impl HeapSize for PageType {
216    fn heap_size(&self) -> usize {
217        0 // no heap allocations
218    }
219}
220
221impl HeapSize for Statistics {
222    fn heap_size(&self) -> usize {
223        match self {
224            Statistics::Boolean(value_statistics) => value_statistics.heap_size(),
225            Statistics::Int32(value_statistics) => value_statistics.heap_size(),
226            Statistics::Int64(value_statistics) => value_statistics.heap_size(),
227            Statistics::Int96(value_statistics) => value_statistics.heap_size(),
228            Statistics::Float(value_statistics) => value_statistics.heap_size(),
229            Statistics::Double(value_statistics) => value_statistics.heap_size(),
230            Statistics::ByteArray(value_statistics) => value_statistics.heap_size(),
231            Statistics::FixedLenByteArray(value_statistics) => value_statistics.heap_size(),
232        }
233    }
234}
235
236impl HeapSize for OffsetIndexMetaData {
237    fn heap_size(&self) -> usize {
238        self.page_locations.heap_size() + self.unencoded_byte_array_data_bytes.heap_size()
239    }
240}
241
242impl HeapSize for ColumnIndexMetaData {
243    fn heap_size(&self) -> usize {
244        match self {
245            Self::NONE => 0,
246            Self::BOOLEAN(native_index) => native_index.heap_size(),
247            Self::INT32(native_index) => native_index.heap_size(),
248            Self::INT64(native_index) => native_index.heap_size(),
249            Self::INT96(native_index) => native_index.heap_size(),
250            Self::FLOAT(native_index) => native_index.heap_size(),
251            Self::DOUBLE(native_index) => native_index.heap_size(),
252            Self::BYTE_ARRAY(native_index) => native_index.heap_size(),
253            Self::FIXED_LEN_BYTE_ARRAY(native_index) => native_index.heap_size(),
254        }
255    }
256}
257
258impl HeapSize for ColumnIndex {
259    fn heap_size(&self) -> usize {
260        self.null_pages.heap_size()
261            + self.boundary_order.heap_size()
262            + self.null_counts.heap_size()
263            + self.definition_level_histograms.heap_size()
264            + self.repetition_level_histograms.heap_size()
265    }
266}
267
268impl<T: ParquetValueType> HeapSize for PrimitiveColumnIndex<T> {
269    fn heap_size(&self) -> usize {
270        self.column_index.heap_size() + self.min_values.heap_size() + self.max_values.heap_size()
271    }
272}
273
274impl HeapSize for ByteArrayColumnIndex {
275    fn heap_size(&self) -> usize {
276        self.column_index.heap_size()
277            + self.min_bytes.heap_size()
278            + self.min_offsets.heap_size()
279            + self.max_bytes.heap_size()
280            + self.max_offsets.heap_size()
281    }
282}
283
284impl<T: ParquetValueType> HeapSize for ValueStatistics<T> {
285    fn heap_size(&self) -> usize {
286        self.min_opt().map(T::heap_size).unwrap_or(0)
287            + self.max_opt().map(T::heap_size).unwrap_or(0)
288    }
289}
290impl HeapSize for bool {
291    fn heap_size(&self) -> usize {
292        0 // no heap allocations
293    }
294}
295impl HeapSize for u8 {
296    fn heap_size(&self) -> usize {
297        0 // no heap allocations
298    }
299}
300impl HeapSize for i32 {
301    fn heap_size(&self) -> usize {
302        0 // no heap allocations
303    }
304}
305impl HeapSize for i64 {
306    fn heap_size(&self) -> usize {
307        0 // no heap allocations
308    }
309}
310
311impl HeapSize for f32 {
312    fn heap_size(&self) -> usize {
313        0 // no heap allocations
314    }
315}
316impl HeapSize for f64 {
317    fn heap_size(&self) -> usize {
318        0 // no heap allocations
319    }
320}
321
322impl HeapSize for usize {
323    fn heap_size(&self) -> usize {
324        0 // no heap allocations
325    }
326}
327
328impl HeapSize for BoundaryOrder {
329    fn heap_size(&self) -> usize {
330        0 // no heap allocations
331    }
332}
333
334impl HeapSize for PageLocation {
335    fn heap_size(&self) -> usize {
336        0 // no heap allocations
337    }
338}
339
340impl HeapSize for ColumnOrder {
341    fn heap_size(&self) -> usize {
342        0 // no heap allocations in ColumnOrder
343    }
344}
parquet/file/metadata/memory.rs

parquet/file/metadata/
memory.rs
