cmd/sort: add random sort order

Author: pjsier

src/cmd/sort.rs

@@ -1,9 +1,11 @@
 use std::cmp;
 
 use CliResult;
+use byteorder::{ByteOrder, LittleEndian};
 use config::{Config, Delimiter};
 use select::SelectColumns;
 use util;
+use rand::{Rng, SeedableRng, StdRng};
 use std::str::from_utf8;
 
 use self::Number::{Float, Int};
@@ -21,6 +23,8 @@ sort options:
                            See 'xsv select --help' for the format details.
     -N, --numeric          Compare according to string numerical value
     -R, --reverse          Reverse order
+    --random               Random order
+    --seed <number>        RNG seed
 
 Common options:
     -h, --help             Display this message
@@ -39,6 +43,8 @@ struct Args {
     flag_select: SelectColumns,
     flag_numeric: bool,
     flag_reverse: bool,
+    flag_random: bool,
+    flag_seed: Option<usize>,
     flag_output: Option<String>,
     flag_no_headers: bool,
     flag_delimiter: Option<Delimiter>,
@@ -48,6 +54,7 @@ pub fn run(argv: &[&str]) -> CliResult<()> {
     let args: Args = util::get_args(USAGE, argv)?;
     let numeric = args.flag_numeric;
     let reverse = args.flag_reverse;
+    let random = args.flag_random;
     let rconfig = Config::new(&args.arg_input)
         .delimiter(args.flag_delimiter)
         .no_headers(args.flag_no_headers)
@@ -58,27 +65,42 @@ pub fn run(argv: &[&str]) -> CliResult<()> {
     let headers = rdr.byte_headers()?.clone();
     let sel = rconfig.selection(&headers)?;
 
+    // Seeding rng
+    let seed = args.flag_seed;
+    let mut rng: StdRng = match seed {
+        None => {
+            StdRng::from_rng(rand::thread_rng()).unwrap()
+        }
+        Some(seed) => {
+            let mut buf = [0u8; 32];
+            LittleEndian::write_u64(&mut buf, seed as u64);
+            SeedableRng::from_seed(buf)
+        }
+    };
+
     let mut all = rdr.byte_records().collect::<Result<Vec<_>, _>>()?;
-    match (numeric, reverse) {
-        (false, false) =>
+    match (numeric, reverse, random) {
+        (_, _, true) =>
+            rng.shuffle(&mut all),
+        (false, false, false) =>
             all.sort_by(|r1, r2| {
                 let a = sel.select(r1);
                 let b = sel.select(r2);
                 iter_cmp(a, b)
             }),
-        (true, false) =>
+        (true, false, false) =>
             all.sort_by(|r1, r2| {
                 let a = sel.select(r1);
                 let b = sel.select(r2);
                 iter_cmp_num(a, b)
             }),
-        (false, true) =>
+        (false, true, false) =>
             all.sort_by(|r1, r2| {
                 let a = sel.select(r1);
                 let b = sel.select(r2);
                 iter_cmp(b, a)
             }),
-        (true, true) =>
+        (true, true, false) =>
             all.sort_by(|r1, r2| {
                 let a = sel.select(r1);
                 let b = sel.select(r2);

tests/test_sort.rs

@@ -128,6 +128,27 @@ fn sort_reverse() {
     assert_eq!(got, expected);
 }
 
+#[test]
+fn sort_random() {
+    let wrk = Workdir::new("sort_random");
+    wrk.create("in.csv", vec![
+        svec!["R", "S"],
+        svec!["1", "b"],
+        svec!["2", "a"],
+    ]);
+
+    let mut cmd = wrk.command("sort");
+    cmd.arg("--random").args(&["--seed", "3"]).arg("in.csv");
+
+    let got: Vec<Vec<String>> = wrk.read_stdout(&mut cmd);
+    let expected = vec![
+        svec!["R", "S"],
+        svec!["2", "a"],
+        svec!["1", "b"],
+    ];
+    assert_eq!(got, expected);
+}
+
 /// Order `a` and `b` lexicographically using `Ord`
 pub fn iter_cmp<A, L, R>(mut a: L, mut b: R) -> cmp::Ordering
         where A: Ord, L: Iterator<Item=A>, R: Iterator<Item=A> {