What are common SQL mistakes in production?

Most SQL bugs are not syntax errors. They are semantic mistakes that return plausible but wrong results, or performance mistakes that only appear at scale. Reviewing queries for these patterns catches many production incidents early.

SELECT *
FROM orders o, customers c
WHERE o.customer_id = c.id
  AND c.status != 'blocked';

Frequent offenders

• NOT IN with a nullable subquery returning no rows
• Using = NULL instead of IS NULL
• Forgetting a join condition and producing a Cartesian product
• Filtering the right table in WHERE after a LEFT JOIN, turning it into an inner join
• Selecting non-aggregated columns that are not grouped
• Applying functions to indexed columns, which blocks normal index use
• Using deep OFFSET instead of keyset pagination
• Using SELECT * in production paths, wasting I/O and breaking on schema changes
• Over-indexing write-heavy tables or assuming every index is used
• Ignoring transaction scope and holding locks during slow external calls

A useful review habit is to ask three questions: what is the grain of each row, which rows are preserved by each join, and what happens when a column is NULL. Those questions catch many correctness bugs before performance even matters.

SELECT id
FROM users u
WHERE NOT EXISTS (
  SELECT 1
  FROM bans b
  WHERE b.user_id = u.id
);