Why should you care?

Higher isolation reduces anomalies but increases blocking. You pick the lowest level that is still correct.

See it live — copy this example

Run in SQL Server Management Studio (SSMS) or Azure Data Studio.

USE DataVerse;
-- Session settings (run in two SSMS windows to experiment)
SET TRANSACTION ISOLATION LEVEL READ COMMITTED;
BEGIN TRAN;
SELECT Balance FROM dbo.Accounts WHERE AccountId = 1;
-- other session updates AccountId 1 here
SELECT Balance FROM dbo.Accounts WHERE AccountId = 1;
COMMIT;
-- Try again with REPEATABLE READ and compare

What happened?

• READ COMMITTED (default) releases shared locks after each statement, so the second SELECT may see a new committed value.
• REPEATABLE READ keeps those locks longer.

Practice next

1. Open two query windows.
2. Run a long BEGIN TRAN update in window A without commit.
3. SELECT the same row in window B under READ COMMITTED vs READ UNCOMMITTED.
4. SET TRANSACTION ISOLATION LEVEL SNAPSHOT after enabling RCSI (next lessons).
5. Watch blocked queries in Activity Monitor.

Remember

Isolation trades correctness vs concurrency. Default is READ COMMITTED. Test with two sessions, not theory alone.

Report vs OLTP

DataVerse reporting uses snapshot isolation; payments stay stricter.

Outcome: Reports rarely block checkout.