PostgreSQL actually treats every SQL statement as being executed within a transaction. If you do not issue a BEGIN command, then each individual statement has an implicit BEGIN and (if successful) COMMIT wrapped around it.
Does this mean that even select statement will be implicitly wrapped in transaction and there is no way to work with postgres out of transaction scope?
Does this mean that even select statement will be implicitly wrapped in transaction and there is no way to work with postgres out of transaction scope?
Yes.
Transactions aren't just about writes, but also isolation level and what SELECTs can read. For example, at the usual isolation level, a SELECT can't read uncommitted writes. It also can't read the updates done from transactions that commit while the SELECT is running. Otherwise it wouldn't have a coherent view of the database and basically everything would break. For example it would look up in an index, then go get the corresponding row, and ithe row could have changed and no longer correspond to the index.
That requires a way to take a "snapshot" of the database right before the SELECT begins to execute, and that is one important function of the transaction mechanism.
