What methods do databases use to ensure protection against race conditions?

The following methods are used to protect against race conditions in databases:

• Transactions and locks: databases apply locks at the row, page, or table level to prevent simultaneous modification of the same data.
• Transaction isolation levels: (Read Uncommitted, Read Committed, Repeatable Read, Serializable) regulate the visibility of changes between transactions, reducing the likelihood of conflicts.
• Optimistic locking: when updating data, it checks whether the data has changed since it was read (for example, using a version or timestamp). If it has changed, the operation is rolled back.
• Pessimistic locking: data is locked immediately upon reading to prevent changes by other transactions.

Example of optimistic locking using a record version:

// Assume there is a Version field in the table
var entity = dbContext.Entities.Find(id);
var currentVersion = entity.Version;

// Change data
entity.Name = "New Name";

// When saving, check that the version has not changed
try {
    dbContext.SaveChanges();
} catch (DbUpdateConcurrencyException) {
    // Handle conflict — data was changed by another process
}