June 2018 Update

Task/Hash-join/C++/hash-join.cpp

@@ -0,0 +1,64 @@
+#include <iostream>
+#include <string>
+#include <vector>
+#include <unordered_map>
+
+using tab_t = std::vector<std::vector<std::string>>;
+tab_t tab1 {
+// Age  Name
+  {"27", "Jonah"}
+, {"18", "Alan"}
+, {"28", "Glory"}
+, {"18", "Popeye"}
+, {"28", "Alan"}
+};
+
+tab_t tab2 {
+// Character  Nemesis
+  {"Jonah", "Whales"}
+, {"Jonah", "Spiders"}
+, {"Alan", "Ghosts"}
+, {"Alan", "Zombies"}
+, {"Glory", "Buffy"}
+};
+
+std::ostream& operator<<(std::ostream& o, const tab_t& t) {
+  for(size_t i = 0; i < t.size(); ++i) {
+    o << i << ":";
+    for(const auto& e : t[i])
+      o << '\t' << e;
+    o << std::endl;
+  }
+  return o;
+}
+
+tab_t Join(const tab_t& a, size_t columna, const tab_t& b, size_t columnb) {
+  std::unordered_multimap<std::string, size_t> hashmap;
+  // hash
+  for(size_t i = 0; i < a.size(); ++i) {
+    hashmap.insert(std::make_pair(a[i][columna], i));
+  }
+  // map
+  tab_t result;
+  for(size_t i = 0; i < b.size(); ++i) {
+    auto range = hashmap.equal_range(b[i][columnb]);
+    for(auto it = range.first; it != range.second; ++it) {
+      tab_t::value_type row;
+      row.insert(row.end() , a[it->second].begin() , a[it->second].end());
+      row.insert(row.end() , b[i].begin()          , b[i].end());
+      result.push_back(std::move(row));
+    }
+  }
+  return result;
+}
+
+int main(int argc, char const *argv[])
+{
+  using namespace std;
+  int ret = 0;
+  cout << "Table A: "       << endl << tab1 << endl;
+  cout << "Table B: "       << endl << tab2 << endl;
+  auto tab3 = Join(tab1, 1, tab2, 0);
+  cout << "Joined tables: " << endl << tab3 << endl;
+  return ret;
+}

Task/Hash-join/Perl-6/hash-join.pl6

@@ -0,0 +1,27 @@
+sub hash-join(@a, &a, @b, &b) {
+    my %hash := @b.classify(&b);
+
+    @a.map: -> $a {
+        |(%hash{a $a} // next).map: -> $b { [$a, $b] }
+    }
+}
+
+# Testing:
+
+my @A =
+    [27, "Jonah"],
+    [18, "Alan"],
+    [28, "Glory"],
+    [18, "Popeye"],
+    [28, "Alan"],
+;
+
+my @B =
+    ["Jonah", "Whales"],
+    ["Jonah", "Spiders"],
+    ["Alan", "Ghosts"],
+    ["Alan", "Zombies"],
+    ["Glory", "Buffy"],
+;
+
+.say for hash-join @A, *[1], @B, *[0];

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);
     }
 }