June 2018 Update

Task/Hash-join/Rust/hash-join.rust

@@ -1,26 +1,43 @@
 use std::collections::HashMap;
+use std::hash::Hash;
+
+// If you know one of the tables is smaller, it is best to make it the second parameter.
+fn hash_join<A, B, K>(first: &[(K, A)], second: &[(K, B)]) -> Vec<(A, K, B)>
+where
+    K: Hash + Eq + Copy,
+    A: Copy,
+    B: Copy,
+{
+    let mut hash_map = HashMap::new();
+
+    // hash phase
+    for &(key, val_a) in second {
+        // collect all values by their keys, appending new ones to each existing entry
+        hash_map.entry(key).or_insert_with(Vec::new).push(val_a);
+    }
+
+    let mut result = Vec::new();
+    // join phase
+    for &(key, val_b) in first {
+        if let Some(vals) = hash_map.get(&key) {
+            let tuples = vals.iter().map(|&val_a| (val_b, key, val_a));
+            result.extend(tuples);
+        }
+    }
+
+    result
+}
 
 fn main() {
-    let table_a = vec![
-        (27, "Jonah"), (18, "Alan"), (28, "Glory"),
-        (18, "Popeye"), (28, "Alan")
-    ];
-    let table_b = vec![
+    let table1 = [("Jonah", 27), ("Alan", 18), ("Glory", 28), ("Popeye", 18), ("Alan", 28)];
+    let table2 = [
         ("Jonah", "Whales"), ("Jonah", "Spiders"), ("Alan", "Ghosts"),
         ("Alan", "Zombies"), ("Glory", "Buffy")
     ];
-    // hash phase
-    let mut h = HashMap::new();
-    for (i, a) in table_a.iter().enumerate() {
-        h.entry(a.1).or_insert_with(Vec::new).push(i);
-    }
-    // join phase
-    for b in table_b {
-        if let Some(vals) = h.get(b.0) {
-            for &val in vals {
-                let a = table_a.get(val).unwrap();
-                println!("{:?} {:?}", a, b);
-            }
-        }
+    let result = hash_join(&table1, &table2);
+    println!("Age | Character Name | Nemesis");
+    println!("----|----------------|--------");
+    for (age, name, nemesis) in result {
+        println!("{:<3} | {:^14} | {}", age, name, nemesis);
     }
 }