Thread: WITH x AS (...) and visibility in UPDATE
Hello all, I am trying out PostgreSQL 9.1 Beta 3. In particular, I am very interested in WITH x AS (...) construction. drop table if exists t; create table t ( identifier serial, title text ); with c as ( insert into t (title) values ('old') returning * ) update t set title = 'new' from c where t.identifier = c.identifier; select * from t; Can someone explain why this returns 'old' instead of 'new'? Is the new row not yet visible when the update is evaluated? Thanks.
On Wed, Jul 27, 2011 at 3:18 PM, Peter V <peterv861908@hotmail.com> wrote: > > Hello all, > > I am trying out PostgreSQL 9.1 Beta 3. In particular, I am very interested in WITH x AS (...) construction. > > drop table if exists t; > create table t > ( > identifier serial, > title text > ); > > with c as > ( > insert into t (title) values ('old') returning * > ) > update t set title = 'new' from c where t.identifier = c.identifier; > > select * from t; > > Can someone explain why this returns 'old' instead of 'new'? Is the new row not yet visible when the update is evaluated? because the update statement isn't doing anything. (you could have confirmed this by adding 'returning *' to the update. While the 'from c' is working, you can't join back to t yet because the statement hasn't resolved. here's a reduced form of your problem: postgres=# with c as ( insert into t (title) values ('old') returning * ) select * from t join c using (identifier); The join fails because at the time it happens t isn't yet populated. merlin
---------------------------------------- > Date: Wed, 27 Jul 2011 15:58:04 -0500 > Subject: Re: [GENERAL] WITH x AS (...) and visibility in UPDATE > From: mmoncure@gmail.com > To: peterv861908@hotmail.com > CC: pgsql-general@postgresql.org > > On Wed, Jul 27, 2011 at 3:18 PM, Peter V <peterv861908@hotmail.com> wrote: > > > > Hello all, > > > > I am trying out PostgreSQL 9.1 Beta 3. In particular, I am very interested in WITH x AS (...) construction. > > > > drop table if exists t; > > create table t > > ( > > identifier serial, > > title text > > ); > > > > with c as > > ( > > insert into t (title) values ('old') returning * > > ) > > update t set title = 'new' from c where t.identifier = c.identifier; > > > > select * from t; > > > > Can someone explain why this returns 'old' instead of 'new'? Is the new row not yet visible when the update is evaluated? > > because the update statement isn't doing anything. (you could have > confirmed this by adding 'returning *' to the update. > > While the 'from c' is working, you can't join back to t yet because > the statement hasn't resolved. here's a reduced form of your problem: > > postgres=# with c as > ( > insert into t (title) values ('old') returning * > ) select * from t join c using (identifier); > > The join fails because at the time it happens t isn't yet populated. > > merlin This makes sense. I thought that the insert was evaluated first, before the join is resolved. This isn't the case apparently. Is there another way to force this? That is, without extracting it to two statements ;) Thanks.
On Wed, Jul 27, 2011 at 4:03 PM, Peter V <peterv861908@hotmail.com> wrote: >> On Wed, Jul 27, 2011 at 3:18 PM, Peter V <peterv861908@hotmail.com> wrote: >> > >> > Hello all, >> > >> > I am trying out PostgreSQL 9.1 Beta 3. In particular, I am very interested in WITH x AS (...) construction. >> > >> > drop table if exists t; >> > create table t >> > ( >> > identifier serial, >> > title text >> > ); >> > >> > with c as >> > ( >> > insert into t (title) values ('old') returning * >> > ) >> > update t set title = 'new' from c where t.identifier = c.identifier; >> > >> > select * from t; >> > >> > > Can someone explain why this returns 'old' instead of 'new'? Is > the new row not yet visible when the update is evaluated? >> >> because the update statement isn't doing anything. (you could have >> confirmed this by adding 'returning *' to the update. >> >> While the 'from c' is working, you can't join back to t yet because >> the statement hasn't resolved. here's a reduced form of your problem: >> >> postgres=# with c as >> ( >> insert into t (title) values ('old') returning * >> ) select * from t join c using (identifier); >> >> The join fails because at the time it happens t isn't yet populated. >> >> merlin > > > > This makes sense. I thought that the insert was evaluated first, before the join is resolved. This isn't the case apparently. > > Is there another way to force this? That is, without extracting it to two statements ;) not in the exact sense you were trying. what is it you are trying to do in general? I'm having trouble understanding your use-case. merlin
---------------------------------------- > Date: Wed, 27 Jul 2011 16:16:48 -0500 > Subject: Re: [GENERAL] WITH x AS (...) and visibility in UPDATE > From: mmoncure@gmail.com > To: peterv861908@hotmail.com > CC: pgsql-general@postgresql.org > > On Wed, Jul 27, 2011 at 4:03 PM, Peter V <peterv861908@hotmail.com> wrote: > >> On Wed, Jul 27, 2011 at 3:18 PM, Peter V <peterv861908@hotmail.com> wrote: > >> > > >> > Hello all, > >> > > >> > I am trying out PostgreSQL 9.1 Beta 3. In particular, I am very interested in WITH x AS (...) construction. > >> > > >> > drop table if exists t; > >> > create table t > >> > ( > >> > identifier serial, > >> > title text > >> > ); > >> > > >> > with c as > >> > ( > >> > insert into t (title) values ('old') returning * > >> > ) > >> > update t set title = 'new' from c where t.identifier = c.identifier; > >> > > >> > select * from t; > >> > > >> > > > Can someone explain why this returns 'old' instead of 'new'? Is > > the new row not yet visible when the update is evaluated? > >> > >> because the update statement isn't doing anything. (you could have > >> confirmed this by adding 'returning *' to the update. > >> > >> While the 'from c' is working, you can't join back to t yet because > >> the statement hasn't resolved. here's a reduced form of your problem: > >> > >> postgres=# with c as > >> ( > >> insert into t (title) values ('old') returning * > >> ) select * from t join c using (identifier); > >> > >> The join fails because at the time it happens t isn't yet populated. > >> > >> merlin > > > > > > > > This makes sense. I thought that the insert was evaluated first, before the join is resolved. This isn't the case apparently. > > > > Is there another way to force this? That is, without extracting it to two statements ;) > > not in the exact sense you were trying. what is it you are trying to > do in general? I'm having trouble understanding your use-case. > I want to apply updates on a copy of a row, instead on the row itself. The queries are above were simplied to demonstratethe problem. So basically I want to do: 1) create the copy of the row and return the identifier 2) apply updates on the new row identified by the identifier returned in step 1 If possible, I want to write this in a single command, to avoid overhead and mistakes. I tried writing a rewrite rule or before trigger, but it becomes quickly a mess to avoid infinite loops. Any ideas are welcome. Thanks.
On 7/27/2011 4:22 PM, Peter V wrote: > I want to apply updates on a copy of a row, instead on the row itself. > The queries are above were simplied to demonstrate the problem. > So basically I want to do: > > 1) create the copy of the row and return the identifier > 2) apply updates on the new row identified by the identifier returned in step 1 > > If possible, I want to write this in a single command, to avoid overhead and mistakes. > > I tried writing a rewrite rule or before trigger, but it becomes quickly a mess to avoid infinite loops. > > Any ideas are welcome. Thanks. > Maybe I'm totally missing something, but why insert a copy and then update instead of directly insert a mutated copy? Something like: INSERT INTO t (foo, bar) SELECT 'my new foo', t.bar FROM t WHERE id=123; Wouldn't the above construction let you make a new row with some new values and some copied values? -- Jack Christensen jackc@hylesanderson.edu
> 1) create the copy of the row and return the identifier > 2) apply updates on the new row identified by the identifier returned in step 1 > > If possible, I want to write this in a single command, to avoid overhead and mistakes. > > I tried writing a rewrite rule or before trigger, but it becomes quickly a mess to avoid infinite loops. > > Any ideas are welcome. Thanks. > > Insert into table (cols) Select ... From table; As mentioned previously just modify the original row before inserting it into the table as a new record. David J.