use hashing for classify and deduplicate

src/algorithm/monadic.rs

@@ -2,7 +2,7 @@
 
 use std::{
  cmp::Ordering,
- collections::{BTreeMap, BTreeSet, HashMap},
+ collections::{HashMap, HashSet},
  f64::consts::{PI, TAU},
  iter::repeat,
  mem::size_of,
@@ -636,11 +636,11 @@ impl<T: ArrayValue> Array<T> {
  if self.rank() == 0 {
  return Err(env.error("Cannot classify a rank-0 array"));
  }
- let mut classes = BTreeMap::new();
+ let mut classes = HashMap::new();
  let mut classified = Vec::with_capacity(self.row_count());
- for row in self.rows() {
+ for row in self.row_slices() {
  let new_class = classes.len();
- let class = *classes.entry(row).or_insert(new_class);
+ let class = *classes.entry(ArrayCmpSlice(row)).or_insert(new_class);
  classified.push(class);
  }
  Ok(classified)
@@ -651,7 +651,7 @@ impl<T: ArrayValue> Array<T> {
  return;
  }
  let mut deduped = CowSlice::new();
- let mut seen = BTreeSet::new();
+ let mut seen = HashSet::new();
  let mut new_len = 0;
  for row in self.row_slices() {
  if seen.insert(ArrayCmpSlice(row)) {

tests/units.ua

@@ -173,6 +173,17 @@ ParseOrZero ← ⍣⋕⋅⋅0
 ⍤⊃⋅∘≍ 1 ⊗ 5 [1 5 5]
 ⍤⊃⋅∘≍ [1] ⊗ [5] [1 5 5]
 
+# Classify
+⍤⊃⋅∘≍ [0 1 2 3 4] ⊛ [2 9 4 8 3]
+⍤⊃⋅∘≍ [0 1 2 3 4] ⊛ ↯5_2⇡10
+⍤⊃⋅∘≍ [0 1 0 1 0] ⊛ ↯5_2⇡4
+
+# Deduplicate
+⍤⊃⋅∘≍ [1 2 3 4 5] ⊝ [1 2 3 4 5]
+⍤⊃⋅∘≍ [1 2 3 4 5] ⊝ [1 2 3 4 5 5 5 5 5 5]
+⍤⊃⋅∘≍ [4 8 2 9 1 3] ⊝ [4 8 2 9 1 8 3 9 4 9 2 8]
+⍤⊃⋅∘≍ [1_2 3_4 5_6] ⊝ [1_2 3_4 5_6 3_4]
+
 # Bits
 ⍤⊃⋅∘≍ ⍜⋯∘ .5
 ⍤⊃⋅∘≍ ⍜⋯∘ .15