Tell me about locks in Postgres.

In PostgreSQL, locks are used to ensure data integrity and consistency when multiple transactions access the database concurrently.

Main types of locks:

• Row-level locks — applied to specific rows in a table, allowing other transactions to work with different rows without conflicts. For example, SELECT ... FOR UPDATE places an exclusive lock on selected rows.

• Table-level locks — lock the entire table for certain operations. For example, LOCK TABLE explicitly locks a table.

• Advisory locks — user-defined locks that do not affect the database engine's operation but are used for synchronization at the application level.

PostgreSQL uses a multi-level locking system with different modes (AccessShareLock, RowExclusiveLock, etc.), which are compatible or conflict with each other.

Example in Go using database/sql:

import (
    "database/sql"
    _ "github.com/lib/pq"
    "log"
)

func lockRow(db *sql.DB, id int) error {
    tx, err := db.Begin()
    if err != nil {
        return err
    }
    defer tx.Rollback()

    // Lock the row with the specified id for update
    row := tx.QueryRow("SELECT * FROM users WHERE id=$1 FOR UPDATE", id)
    // process row...

    return tx.Commit()
}

It is important to understand that improper use of locks can lead to deadlocks and reduced performance, so transaction logic should be carefully designed.