join: add benchmarks (#10005)

.github/workflows/benchmarks.yml

@@ -31,6 +31,7 @@ jobs:
           - { package: uu_du }
           - { package: uu_expand }
           - { package: uu_fold }
+          - { package: uu_join }
           - { package: uu_ls }
           - { package: uu_mv }
           - { package: uu_nl }

Cargo.lock

@@ -3462,8 +3462,10 @@ name = "uu_join"
 version = "0.5.0"
 dependencies = [
  "clap",
+ "codspeed-divan-compat",
  "fluent",
  "memchr",
+ "tempfile",
  "thiserror 2.0.17",
  "uucore",
 ]

src/uu/join/Cargo.toml

@@ -27,3 +27,12 @@ fluent = { workspace = true }
 [[bin]]
 name = "join"
 path = "src/main.rs"
+
+[dev-dependencies]
+divan = { workspace = true }
+tempfile = { workspace = true }
+uucore = { workspace = true, features = ["benchmark"] }
+
+[[bench]]
+name = "join_bench"
+harness = false

src/uu/join/benches/join_bench.rs

@@ -0,0 +1,115 @@
+// This file is part of the uutils coreutils package.
+//
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use divan::{Bencher, black_box};
+use std::{fs::File, io::Write};
+use tempfile::TempDir;
+use uu_join::uumain;
+use uucore::benchmark::run_util_function;
+
+/// Create two sorted files with matching keys for join benchmarking
+fn create_join_files(temp_dir: &TempDir, num_lines: usize) -> (String, String) {
+    let file1_path = temp_dir.path().join("file1.txt");
+    let file2_path = temp_dir.path().join("file2.txt");
+
+    let mut file1 = File::create(&file1_path).unwrap();
+    let mut file2 = File::create(&file2_path).unwrap();
+
+    for i in 0..num_lines {
+        writeln!(file1, "{i:08} field1_{i} field2_{i}").unwrap();
+        writeln!(file2, "{i:08} data1_{i} data2_{i}").unwrap();
+    }
+
+    (
+        file1_path.to_str().unwrap().to_string(),
+        file2_path.to_str().unwrap().to_string(),
+    )
+}
+
+/// Create two files with partial overlap for join benchmarking
+fn create_partial_overlap_files(
+    temp_dir: &TempDir,
+    num_lines: usize,
+    overlap_ratio: f64,
+) -> (String, String) {
+    let file1_path = temp_dir.path().join("file1.txt");
+    let file2_path = temp_dir.path().join("file2.txt");
+
+    let mut file1 = File::create(&file1_path).unwrap();
+    let mut file2 = File::create(&file2_path).unwrap();
+
+    let overlap_count = (num_lines as f64 * overlap_ratio) as usize;
+
+    // File 1: keys 0 to num_lines-1
+    for i in 0..num_lines {
+        writeln!(file1, "{i:08} f1_data_{i}").unwrap();
+    }
+
+    // File 2: keys (num_lines - overlap_count) to (2*num_lines - overlap_count - 1)
+    let start = num_lines - overlap_count;
+    for i in 0..num_lines {
+        writeln!(file2, "{:08} f2_data_{}", start + i, i).unwrap();
+    }
+
+    (
+        file1_path.to_str().unwrap().to_string(),
+        file2_path.to_str().unwrap().to_string(),
+    )
+}
+
+/// Benchmark basic join with fully matching keys
+#[divan::bench]
+fn join_full_match(bencher: Bencher) {
+    let num_lines = 10000;
+    let temp_dir = TempDir::new().unwrap();
+    let (file1, file2) = create_join_files(&temp_dir, num_lines);
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &[&file1, &file2]));
+    });
+}
+
+/// Benchmark join with partial overlap (50%)
+#[divan::bench]
+fn join_partial_overlap(bencher: Bencher) {
+    let num_lines = 10000;
+    let temp_dir = TempDir::new().unwrap();
+    let (file1, file2) = create_partial_overlap_files(&temp_dir, num_lines, 0.5);
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &[&file1, &file2]));
+    });
+}
+
+/// Benchmark join with custom field separator
+#[divan::bench]
+fn join_custom_separator(bencher: Bencher) {
+    let num_lines = 10000;
+    let temp_dir = TempDir::new().unwrap();
+    let file1_path = temp_dir.path().join("file1.txt");
+    let file2_path = temp_dir.path().join("file2.txt");
+
+    let mut file1 = File::create(&file1_path).unwrap();
+    let mut file2 = File::create(&file2_path).unwrap();
+
+    for i in 0..num_lines {
+        writeln!(file1, "{i:08}\tfield1_{i}\tfield2_{i}").unwrap();
+        writeln!(file2, "{i:08}\tdata1_{i}\tdata2_{i}").unwrap();
+    }
+
+    let file1_str = file1_path.to_str().unwrap();
+    let file2_str = file2_path.to_str().unwrap();
+
+    bencher.bench(|| {
+        black_box(run_util_function(
+            uumain,
+            &["-t", "\t", file1_str, file2_str],
+        ));
+    });
+}
+
+fn main() {
+    divan::main();
+}