|
| 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 | +#[macro_use] |
| 19 | +extern crate criterion; |
| 20 | + |
| 21 | +use criterion::{BenchmarkId, Criterion}; |
| 22 | +use datafusion_expr::lit; |
| 23 | +use datafusion_functions_nested::expr_fn::{ |
| 24 | + array_has, array_has_all, array_has_any, make_array, |
| 25 | +}; |
| 26 | +use std::hint::black_box; |
| 27 | + |
| 28 | +// If not explicitly stated, `array` and `array_size` refer to the haystack array. |
| 29 | +fn criterion_benchmark(c: &mut Criterion) { |
| 30 | + // Test different array sizes |
| 31 | + let array_sizes = vec![1, 10, 100, 1000, 10000]; |
| 32 | + |
| 33 | + for &size in &array_sizes { |
| 34 | + bench_array_has(c, size); |
| 35 | + bench_array_has_all(c, size); |
| 36 | + bench_array_has_any(c, size); |
| 37 | + } |
| 38 | + |
| 39 | + // Specific benchmarks for string arrays (common use case) |
| 40 | + bench_array_has_strings(c); |
| 41 | + bench_array_has_all_strings(c); |
| 42 | + bench_array_has_any_strings(c); |
| 43 | + |
| 44 | + // Edge cases |
| 45 | + bench_array_has_edge_cases(c); |
| 46 | +} |
| 47 | + |
| 48 | +fn bench_array_has(c: &mut Criterion, array_size: usize) { |
| 49 | + let mut group = c.benchmark_group("array_has_i64"); |
| 50 | + |
| 51 | + // Benchmark: element found at beginning |
| 52 | + group.bench_with_input( |
| 53 | + BenchmarkId::new("found_at_start", array_size), |
| 54 | + &array_size, |
| 55 | + |b, &size| { |
| 56 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 57 | + let list_array = make_array(array); |
| 58 | + let needle = lit(0_i64); |
| 59 | + |
| 60 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 61 | + }, |
| 62 | + ); |
| 63 | + |
| 64 | + // Benchmark: element found at end |
| 65 | + group.bench_with_input( |
| 66 | + BenchmarkId::new("found_at_end", array_size), |
| 67 | + &array_size, |
| 68 | + |b, &size| { |
| 69 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 70 | + let list_array = make_array(array); |
| 71 | + let needle = lit((size - 1) as i64); |
| 72 | + |
| 73 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 74 | + }, |
| 75 | + ); |
| 76 | + |
| 77 | + // Benchmark: element not found |
| 78 | + group.bench_with_input( |
| 79 | + BenchmarkId::new("not_found", array_size), |
| 80 | + &array_size, |
| 81 | + |b, &size| { |
| 82 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 83 | + let list_array = make_array(array); |
| 84 | + let needle = lit(-1_i64); // Not in array |
| 85 | + |
| 86 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 87 | + }, |
| 88 | + ); |
| 89 | + |
| 90 | + group.finish(); |
| 91 | +} |
| 92 | + |
| 93 | +fn bench_array_has_all(c: &mut Criterion, array_size: usize) { |
| 94 | + let mut group = c.benchmark_group("array_has_all"); |
| 95 | + |
| 96 | + // Benchmark: all elements found (small needle) |
| 97 | + group.bench_with_input( |
| 98 | + BenchmarkId::new("all_found_small_needle", array_size), |
| 99 | + &array_size, |
| 100 | + |b, &size| { |
| 101 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 102 | + let list_array = make_array(array); |
| 103 | + let needle_array = make_array(vec![lit(0_i64), lit(1_i64), lit(2_i64)]); |
| 104 | + |
| 105 | + b.iter(|| black_box(array_has_all(list_array.clone(), needle_array.clone()))) |
| 106 | + }, |
| 107 | + ); |
| 108 | + |
| 109 | + // Benchmark: all elements found (medium needle - 10% of haystack) |
| 110 | + group.bench_with_input( |
| 111 | + BenchmarkId::new("all_found_medium_needle", array_size), |
| 112 | + &array_size, |
| 113 | + |b, &size| { |
| 114 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 115 | + let list_array = make_array(array); |
| 116 | + let needle_size = (size / 10).max(1); |
| 117 | + let needle = (0..needle_size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 118 | + let needle_array = make_array(needle); |
| 119 | + |
| 120 | + b.iter(|| black_box(array_has_all(list_array.clone(), needle_array.clone()))) |
| 121 | + }, |
| 122 | + ); |
| 123 | + |
| 124 | + // Benchmark: not all found (early exit) |
| 125 | + group.bench_with_input( |
| 126 | + BenchmarkId::new("early_exit", array_size), |
| 127 | + &array_size, |
| 128 | + |b, &size| { |
| 129 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 130 | + let list_array = make_array(array); |
| 131 | + let needle_array = make_array(vec![lit(0_i64), lit(-1_i64)]); // -1 not in array |
| 132 | + |
| 133 | + b.iter(|| black_box(array_has_all(list_array.clone(), needle_array.clone()))) |
| 134 | + }, |
| 135 | + ); |
| 136 | + |
| 137 | + group.finish(); |
| 138 | +} |
| 139 | + |
| 140 | +fn bench_array_has_any(c: &mut Criterion, array_size: usize) { |
| 141 | + let mut group = c.benchmark_group("array_has_any"); |
| 142 | + |
| 143 | + // Benchmark: first element matches (best case) |
| 144 | + group.bench_with_input( |
| 145 | + BenchmarkId::new("first_match", array_size), |
| 146 | + &array_size, |
| 147 | + |b, &size| { |
| 148 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 149 | + let list_array = make_array(array); |
| 150 | + let needle_array = make_array(vec![lit(0_i64), lit(-1_i64), lit(-2_i64)]); |
| 151 | + |
| 152 | + b.iter(|| black_box(array_has_any(list_array.clone(), needle_array.clone()))) |
| 153 | + }, |
| 154 | + ); |
| 155 | + |
| 156 | + // Benchmark: last element matches (worst case) |
| 157 | + group.bench_with_input( |
| 158 | + BenchmarkId::new("last_match", array_size), |
| 159 | + &array_size, |
| 160 | + |b, &size| { |
| 161 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 162 | + let list_array = make_array(array); |
| 163 | + let needle_array = make_array(vec![lit(-1_i64), lit(-2_i64), lit(0_i64)]); |
| 164 | + |
| 165 | + b.iter(|| black_box(array_has_any(list_array.clone(), needle_array.clone()))) |
| 166 | + }, |
| 167 | + ); |
| 168 | + |
| 169 | + // Benchmark: no match |
| 170 | + group.bench_with_input( |
| 171 | + BenchmarkId::new("no_match", array_size), |
| 172 | + &array_size, |
| 173 | + |b, &size| { |
| 174 | + let array = (0..size).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 175 | + let list_array = make_array(array); |
| 176 | + let needle_array = make_array(vec![lit(-1_i64), lit(-2_i64), lit(-3_i64)]); |
| 177 | + |
| 178 | + b.iter(|| black_box(array_has_any(list_array.clone(), needle_array.clone()))) |
| 179 | + }, |
| 180 | + ); |
| 181 | + |
| 182 | + group.finish(); |
| 183 | +} |
| 184 | + |
| 185 | +fn bench_array_has_strings(c: &mut Criterion) { |
| 186 | + let mut group = c.benchmark_group("array_has_strings"); |
| 187 | + |
| 188 | + // Benchmark with string arrays (common use case for tickers, tags, etc.) |
| 189 | + let sizes = vec![10, 100, 1000]; |
| 190 | + |
| 191 | + for &size in &sizes { |
| 192 | + group.bench_with_input(BenchmarkId::new("found", size), &size, |b, &size| { |
| 193 | + let array = (0..size) |
| 194 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 195 | + .collect::<Vec<_>>(); |
| 196 | + let list_array = make_array(array); |
| 197 | + let needle = lit("TICKER0005"); |
| 198 | + |
| 199 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 200 | + }); |
| 201 | + |
| 202 | + group.bench_with_input(BenchmarkId::new("not_found", size), &size, |b, &size| { |
| 203 | + let array = (0..size) |
| 204 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 205 | + .collect::<Vec<_>>(); |
| 206 | + let list_array = make_array(array); |
| 207 | + let needle = lit("NOTFOUND"); |
| 208 | + |
| 209 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 210 | + }); |
| 211 | + } |
| 212 | + |
| 213 | + group.finish(); |
| 214 | +} |
| 215 | + |
| 216 | +fn bench_array_has_all_strings(c: &mut Criterion) { |
| 217 | + let mut group = c.benchmark_group("array_has_all_strings"); |
| 218 | + |
| 219 | + // Realistic scenario: checking if a portfolio contains certain tickers |
| 220 | + let portfolio_size = 100; |
| 221 | + let check_sizes = vec![1, 3, 5, 10]; |
| 222 | + |
| 223 | + for &check_size in &check_sizes { |
| 224 | + group.bench_with_input( |
| 225 | + BenchmarkId::new("all_found", check_size), |
| 226 | + &check_size, |
| 227 | + |b, &check_size| { |
| 228 | + let portfolio = (0..portfolio_size) |
| 229 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 230 | + .collect::<Vec<_>>(); |
| 231 | + let list_array = make_array(portfolio); |
| 232 | + |
| 233 | + let checking = (0..check_size) |
| 234 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 235 | + .collect::<Vec<_>>(); |
| 236 | + let needle_array = make_array(checking); |
| 237 | + |
| 238 | + b.iter(|| { |
| 239 | + black_box(array_has_all(list_array.clone(), needle_array.clone())) |
| 240 | + }) |
| 241 | + }, |
| 242 | + ); |
| 243 | + |
| 244 | + group.bench_with_input( |
| 245 | + BenchmarkId::new("some_missing", check_size), |
| 246 | + &check_size, |
| 247 | + |b, &check_size| { |
| 248 | + let portfolio = (0..portfolio_size) |
| 249 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 250 | + .collect::<Vec<_>>(); |
| 251 | + let list_array = make_array(portfolio); |
| 252 | + |
| 253 | + let mut checking = (0..check_size - 1) |
| 254 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 255 | + .collect::<Vec<_>>(); |
| 256 | + checking.push(lit("NOTFOUND".to_string())); |
| 257 | + let needle_array = make_array(checking); |
| 258 | + |
| 259 | + b.iter(|| { |
| 260 | + black_box(array_has_all(list_array.clone(), needle_array.clone())) |
| 261 | + }) |
| 262 | + }, |
| 263 | + ); |
| 264 | + } |
| 265 | + |
| 266 | + group.finish(); |
| 267 | +} |
| 268 | + |
| 269 | +fn bench_array_has_any_strings(c: &mut Criterion) { |
| 270 | + let mut group = c.benchmark_group("array_has_any_strings"); |
| 271 | + |
| 272 | + let portfolio_size = 100; |
| 273 | + let check_sizes = vec![1, 3, 5, 10]; |
| 274 | + |
| 275 | + for &check_size in &check_sizes { |
| 276 | + group.bench_with_input( |
| 277 | + BenchmarkId::new("first_matches", check_size), |
| 278 | + &check_size, |
| 279 | + |b, &check_size| { |
| 280 | + let portfolio = (0..portfolio_size) |
| 281 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 282 | + .collect::<Vec<_>>(); |
| 283 | + let list_array = make_array(portfolio); |
| 284 | + |
| 285 | + let mut checking = vec![lit("TICKER0000".to_string())]; |
| 286 | + checking.extend((1..check_size).map(|_| lit("NOTFOUND".to_string()))); |
| 287 | + let needle_array = make_array(checking); |
| 288 | + |
| 289 | + b.iter(|| { |
| 290 | + black_box(array_has_any(list_array.clone(), needle_array.clone())) |
| 291 | + }) |
| 292 | + }, |
| 293 | + ); |
| 294 | + |
| 295 | + group.bench_with_input( |
| 296 | + BenchmarkId::new("none_match", check_size), |
| 297 | + &check_size, |
| 298 | + |b, &check_size| { |
| 299 | + let portfolio = (0..portfolio_size) |
| 300 | + .map(|i| lit(format!("TICKER{i:04}"))) |
| 301 | + .collect::<Vec<_>>(); |
| 302 | + let list_array = make_array(portfolio); |
| 303 | + |
| 304 | + let checking = (0..check_size) |
| 305 | + .map(|i| lit(format!("NOTFOUND{i}"))) |
| 306 | + .collect::<Vec<_>>(); |
| 307 | + let needle_array = make_array(checking); |
| 308 | + |
| 309 | + b.iter(|| { |
| 310 | + black_box(array_has_any(list_array.clone(), needle_array.clone())) |
| 311 | + }) |
| 312 | + }, |
| 313 | + ); |
| 314 | + } |
| 315 | + |
| 316 | + group.finish(); |
| 317 | +} |
| 318 | + |
| 319 | +fn bench_array_has_edge_cases(c: &mut Criterion) { |
| 320 | + let mut group = c.benchmark_group("array_has_edge_cases"); |
| 321 | + |
| 322 | + // Empty array |
| 323 | + group.bench_function("empty_array", |b| { |
| 324 | + let list_array = make_array(vec![]); |
| 325 | + let needle = lit(1_i64); |
| 326 | + |
| 327 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 328 | + }); |
| 329 | + |
| 330 | + // Single element array - found |
| 331 | + group.bench_function("single_element_found", |b| { |
| 332 | + let list_array = make_array(vec![lit(1_i64)]); |
| 333 | + let needle = lit(1_i64); |
| 334 | + |
| 335 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 336 | + }); |
| 337 | + |
| 338 | + // Single element array - not found |
| 339 | + group.bench_function("single_element_not_found", |b| { |
| 340 | + let list_array = make_array(vec![lit(1_i64)]); |
| 341 | + let needle = lit(2_i64); |
| 342 | + |
| 343 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 344 | + }); |
| 345 | + |
| 346 | + // Array with duplicates |
| 347 | + group.bench_function("array_with_duplicates", |b| { |
| 348 | + let array = vec![lit(1_i64); 1000]; |
| 349 | + let list_array = make_array(array); |
| 350 | + let needle = lit(1_i64); |
| 351 | + |
| 352 | + b.iter(|| black_box(array_has(list_array.clone(), needle.clone()))) |
| 353 | + }); |
| 354 | + |
| 355 | + // array_has_all: empty needle |
| 356 | + group.bench_function("array_has_all_empty_needle", |b| { |
| 357 | + let array = (0..1000).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 358 | + let list_array = make_array(array); |
| 359 | + let needle_array = make_array(vec![]); |
| 360 | + |
| 361 | + b.iter(|| black_box(array_has_all(list_array.clone(), needle_array.clone()))) |
| 362 | + }); |
| 363 | + |
| 364 | + // array_has_any: empty needle |
| 365 | + group.bench_function("array_has_any_empty_needle", |b| { |
| 366 | + let array = (0..1000).map(|i| lit(i as i64)).collect::<Vec<_>>(); |
| 367 | + let list_array = make_array(array); |
| 368 | + let needle_array = make_array(vec![]); |
| 369 | + |
| 370 | + b.iter(|| black_box(array_has_any(list_array.clone(), needle_array.clone()))) |
| 371 | + }); |
| 372 | + |
| 373 | + group.finish(); |
| 374 | +} |
| 375 | + |
| 376 | +criterion_group!(benches, criterion_benchmark); |
| 377 | +criterion_main!(benches); |
0 commit comments