Implement insert_or_replace (#45)

Co-authored-by: Dr. Patrick Urbanke <patrick@getml.com>

docs/insert.md

@@ -68,6 +68,58 @@ sqlgen::sqlite::connect("database.db")
     .value();
 ```
 
+### With Replacement
+
+Replace existing rows:
+
+```cpp
+const auto people1 = std::vector<Person>({
+    Person{.id = 0, .first_name = "Homer", .last_name = "Simpson", .age = 45},
+    Person{.id = 1, .first_name = "Bart", .last_name = "Simpson", .age = 10}
+});
+
+const auto people2 = std::vector<Person>({
+    Person{.id = 1, .first_name = "Bartholomew", .last_name = "Simpson", .age = 10}
+});
+
+using namespace sqlgen;
+
+const auto result = sqlite::connect()
+    .and_then(create_table<Person> | if_not_exists)
+    .and_then(insert(std::ref(people1)))
+    .and_then(insert_or_replace(std::ref(people2)))
+    .value();
+```
+
+This generates the following SQL:
+
+```sql
+CREATE TABLE IF NOT EXISTS "Person" (
+    "id" INTEGER PRIMARY KEY,
+    "first_name" TEXT NOT NULL,
+    "last_name" TEXT NOT NULL,
+    "age" INTEGER NOT NULL
+);
+INSERT INTO "Person" ("id", "first_name", "last_name", "age") VALUES (?, ?, ?, ?);
+INSERT INTO "Person" ("id", "first_name", "last_name", "age") VALUES (?, ?, ?, ?)
+ON CONFLICT (id) DO UPDATE SET
+  id=excluded.id,
+  first_name=excluded.first_name,
+  last_name=excluded.last_name,
+  age=excluded.age;
+```
+
+The SQL generated by the PostgreSQL backend for `insert_or_replace(...)` is similar. The MySQL backend generates `ON DUPLICATE KEY UPDATE` instead:
+```sql
+...
+INSERT INTO `Person` (`id`, `first_name`, `last_name`, `age`) VALUES (?, ?, ?, ?)
+ON DUPLICATE KEY UPDATE
+  id=VALUES(id),
+  first_name=VALUES(first_name),
+  last_name=VALUES(last_name),
+  age=VALUES(age);
+```
+
 ## Example: Full Transaction Usage
 
 Here's a complete example showing how to use `insert` within a transaction:
@@ -195,5 +247,4 @@ While both `insert` and `write` can be used to add data to a database, they serv
   4. Takes a connection and a reference wrapper to a container
 - Unlike `write`, `insert` does not create tables automatically - you must create tables separately using `create_table`
 - The insert operation is atomic within a transaction
-- When using reference wrappers (`std::ref`), the data is not copied, which can be more efficient for large datasets
-
+- When using reference wrappers (`std::ref`), the data is not copied, which can be more efficient for large datasets