How to optimize query for Max(Date) in MySQL - mysql

Question

I have this SQL Query:
SELECT company.*, salesorder.lastOrderDate
FROM company
INNER JOIN
(
SELECT companyId, MAX(orderDate) AS lastOrderDate
FROM salesorder
GROUP BY companyId
) salesorder ON salesorder.companyId = company.companyId;
This gives me one extra column at the end of a company master table with their last order date.
Problem is, when analyzing this query, it seems like it's not that efficient:
Is there a way to make this more efficient?
salesorder:
orderId, companyId, orderDate
1 333 2015-01-01
2 555 2016-01-01
3 333 2017-01-01
company
companyId, name
333 Acme
555 Microsoft
Query:
companyId, name, lastOrderDate
333 Acme 2017-01-01
555 Microsoft 2016-01-01
EXPLAIN SELECT:
CREATE TABLE `salesorder` (
`orderId` int(11) NOT NULL,
`companyId` int(11) DEFAULT NULL,
`orderDate` date DEFAULT NULL,
PRIMARY KEY (`orderId`),
UNIQUE KEY `orderId_UNIQUE` (`orderId`) /*!80000 INVISIBLE */,
KEY `testComposite` (`companyId`,`orderDate`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci;
CREATE TABLE `company` (
`companyId` int(11) NOT NULL,
`name` varchar(45) DEFAULT NULL,
PRIMARY KEY (`companyId`),
UNIQUE KEY `companyId_UNIQUE` (`companyId`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci;

Answer 1

It looks like you could simplify the query like this:
SELECT c.*, MAX(o.OrderDate) As lastOrderDate
FROM company c
INNER JOIN salesorder o on o.companyId = c.companyId
GROUP BY <list all company fields here>;
MySql might even let you get away with just c.companyId in the GROUP BY clause, but that's not really standard and not great practice.

Answer 2

Add the composite index with the columns in this order:
INDEX(companyId, orderDate)
Single column indexes are not as efficient (in this query).
Since a PRIMARY KEY is a unique key, do not redundantly declare a UNIQUE key.
With only a few rows in the table, you cannot trust EXPLAIN (and Explain-like output) to say how bad the query will be. Try it with at least a few dozen rows. And provide EXPLAIN FORMAT=JSON SELECT ...
Note that it says "Using index". That says that the subquery in question can be performed entirely inside the index's BTree. This is 'good'. (I presume you did the EXPLAIN after adding my suggested index?)
Your previous image showed a lot of rows; what gives?
I'm still puzzled as to why there are 3 rows in the EXPLAIN and two table scans. Anyway, here is another formulation to try:
SELECT c.*,
( SELECT MAX(orderDate)
FROM salesorder
WHERE companyId = c.companyId
) AS lastOrderDate
FROM company AS c;
(and my INDEX is still important)