Order totals that do not match their items

Four tables of an online shop. customers (id, name, email, city, created_at): the product requires an email, but in this data it may be NULL or empty; city is optional. orders (id, customer_id, total, status, created_at): status is PAID, PENDING or CANCELLED. order_items (id, order_id, product, qty, price): one row per order line, worth qty times price. payments (id, order_id, amount, status): status is SUCCESS or FAILED. Dates are text in the form YYYY-MM-DD. No foreign keys are enforced, so the data may hold mistakes.