Thread: Mistake in statement example
The following documentation comment has been logged on the website: Page: https://www.postgresql.org/docs/15/transaction-iso.html Description: I believe there is a mistake in an example on https://www.postgresql.org/docs/current/transaction-iso.html section 13.2.1: BEGIN; UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; UPDATE accounts SET balance = balance - 100.00 WHERE acctnum = 7534; COMMIT; The acctnum is expected to be 12345 in both cases.
PG Doc comments form <noreply@postgresql.org> writes: > I believe there is a mistake in an example on > https://www.postgresql.org/docs/current/transaction-iso.html section > 13.2.1: > BEGIN; > UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; > UPDATE accounts SET balance = balance - 100.00 WHERE acctnum = 7534; > COMMIT; > The acctnum is expected to be 12345 in both cases. No, I think that's intentional: the example depicts transferring $100 from account 7534 to account 12345. regards, tom lane
On Wed, Mar 1, 2023 at 9:34 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
PG Doc comments form <noreply@postgresql.org> writes:
> I believe there is a mistake in an example on
> https://www.postgresql.org/docs/current/transaction-iso.html section
> 13.2.1:
> BEGIN;
> UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345;
> UPDATE accounts SET balance = balance - 100.00 WHERE acctnum = 7534;
> COMMIT;
> The acctnum is expected to be 12345 in both cases.
No, I think that's intentional: the example depicts transferring
$100 from account 7534 to account 12345.
That may be, but the descriptive text and point of the example (which isn't atomicity, but concurrency) doesn't even require the second update command to be present. What the example could use is a more traditional two-session depiction of the commands instead of having a single transaction and letting the user envision the correct concurrency.
Something like:
S1: SELECT balance FROM accounts WHERE acctnum = 12345; //100
S1: BEGIN;
S2: BEGIN;
S1: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; //200
S2: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; //WAITING ON S1
S1: COMMIT;
S2: UPDATED; balance = 300
S2: COMMIT;
Though maybe "balance" isn't a good example domain, the incrementing example used just after this one seems more appropriate along with the added benefit of consistency.
David J.
On Wed, Mar 1, 2023 at 09:45:00AM -0700, David G. Johnston wrote: > On Wed, Mar 1, 2023 at 9:34 AM Tom Lane <tgl@sss.pgh.pa.us> wrote: > > PG Doc comments form <noreply@postgresql.org> writes: > > I believe there is a mistake in an example on > > https://www.postgresql.org/docs/current/transaction-iso.html section > > 13.2.1: > > BEGIN; > > UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; > > UPDATE accounts SET balance = balance - 100.00 WHERE acctnum = 7534; > > COMMIT; > > > The acctnum is expected to be 12345 in both cases. > > No, I think that's intentional: the example depicts transferring > $100 from account 7534 to account 12345. > > > > That may be, but the descriptive text and point of the example (which isn't > atomicity, but concurrency) doesn't even require the second update command to > be present. What the example could use is a more traditional two-session > depiction of the commands instead of having a single transaction and letting > the user envision the correct concurrency. > > Something like: > > S1: SELECT balance FROM accounts WHERE acctnum = 12345; //100 > S1: BEGIN; > S2: BEGIN; > S1: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; //200 > S2: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; // > WAITING ON S1 > S1: COMMIT; > S2: UPDATED; balance = 300 > S2: COMMIT; > > Though maybe "balance" isn't a good example domain, the incrementing example > used just after this one seems more appropriate along with the added benefit of > consistency. I developed the attached patch. I explained the example, I mentioned a "second" transaciton, I changed the account number so I can talk about the second statement, because read committed changes the row visibility of the non-first statements, and I changed "transaction" to "statement". -- Bruce Momjian <bruce@momjian.us> https://momjian.us EDB https://enterprisedb.com Only you can decide what is important to you.
Attachment
On Wed, Sep 27, 2023 at 07:23:09PM -0400, Bruce Momjian wrote: > On Wed, Mar 1, 2023 at 09:45:00AM -0700, David G. Johnston wrote: > > That may be, but the descriptive text and point of the example (which isn't > > atomicity, but concurrency) doesn't even require the second update command to > > be present. What the example could use is a more traditional two-session > > depiction of the commands instead of having a single transaction and letting > > the user envision the correct concurrency. > > > > Something like: > > > > S1: SELECT balance FROM accounts WHERE acctnum = 12345; //100 > > S1: BEGIN; > > S2: BEGIN; > > S1: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; //200 > > S2: UPDATE accounts SET balance = balance + 100.00 WHERE acctnum = 12345; // > > WAITING ON S1 > > S1: COMMIT; > > S2: UPDATED; balance = 300 > > S2: COMMIT; > > > > Though maybe "balance" isn't a good example domain, the incrementing example > > used just after this one seems more appropriate along with the added benefit of > > consistency. > > I developed the attached patch. I explained the example, I mentioned a > "second" transaciton, I changed the account number so I can talk about > the second statement, because read committed changes the row visibility > of the non-first statements, and I changed "transaction" to "statement". Patch from September 2023 applied. -- Bruce Momjian <bruce@momjian.us> https://momjian.us EDB https://enterprisedb.com When a patient asks the doctor, "Am I going to die?", he means "Am I going to die soon?"