Problem with aliasing

Started by Jean-Christian Imbeaultabout 23 years ago5 messagesgeneral
Jump to latest
#1Jean-Christian Imbeault
jc@mega-bucks.co.jp

I wrote a plpgsql function to return the maximum of three dates.

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as ma1
from invoices;
ma1
------------
2003-02-25
2003-02-25
(2 rows)

However I'm having problems with the following query:

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as max
from invoices where max <= now();
ERROR: Attribute "max" not found

Why can't the where part of the query see "max"? I've tried aliasing the
returned value but that did not work either:

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as ma1
from invoices where ma1 <= now();
ERROR: Attribute "ma1" not found

What is wrong with my syntax?

Thanks!

Jc

#2Bruce Momjian
bruce@momjian.us
In reply to: Jean-Christian Imbeault (#1)
Re: Problem with aliasing

Jean-Christian Imbeault <jc@mega-bucks.co.jp> writes:

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as ma1 from
invoices where ma1 <= now();
ERROR: Attribute "ma1" not found

What is wrong with my syntax?

Try either of:

SELECT max(req1, req2, req3) AS max FROM invoices WHERE max(req1, req2, req3) <= now()
SELECT * from (select max(req1, req2, req3) AS max FROM invoices) WHERE max <= now()

Whichever seems clearer to you, I think postgres actually runs the two the
same way. Note, you should mark your function immutable so postgres knows it
can optimize the second case into the first.

--
greg

#3Jean-Christian Imbeault
jc@mega-bucks.co.jp
In reply to: Jean-Christian Imbeault (#1)
Re: Problem with aliasing

Greg Stark wrote:

Try either of:

SELECT max(req1, req2, req3) AS max FROM invoices WHERE max(req1, req2, req3) <= now()

Ok, that works. But why can't I alias the result of the max() function
and use the alias in the where clause? Something like:

SELECT max(r1,r2,r3) as max from invoices WHERE max <= now();

Why can't postgres see the alias when inside the WHERE clause?

Jc

#4Nigel J. Andrews
nandrews@investsystems.co.uk
In reply to: Bruce Momjian (#2)
Re: Problem with aliasing

On 21 Feb 2003, Greg Stark wrote:

Jean-Christian Imbeault <jc@mega-bucks.co.jp> writes:

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as ma1 from
invoices where ma1 <= now();
ERROR: Attribute "ma1" not found

What is wrong with my syntax?

Try either of:

SELECT max(req1, req2, req3) AS max FROM invoices WHERE max(req1, req2, req3) <= now()
SELECT * from (select max(req1, req2, req3) AS max FROM invoices) WHERE max <= now()

Whichever seems clearer to you, I think postgres actually runs the two the
same way. Note, you should mark your function immutable so postgres knows it
can optimize the second case into the first.

Doesn't:

SELECT max(req1, req2, req3) AS ma1 FROM invoices HAVING ma1 <= now();

work?

Now I'm going to have to go see if I've got that wrong as well...

--
Nigel J. Andrews

#5Nigel J. Andrews
nandrews@investsystems.co.uk
In reply to: Nigel J. Andrews (#4)
Re: Problem with aliasing

On Fri, 21 Feb 2003, Nigel J. Andrews wrote:

On 21 Feb 2003, Greg Stark wrote:

Jean-Christian Imbeault <jc@mega-bucks.co.jp> writes:

TAL=# select max(req_del_date1, req_del_date2, req_del_date3) as ma1 from
invoices where ma1 <= now();
ERROR: Attribute "ma1" not found

What is wrong with my syntax?

Try either of:

SELECT max(req1, req2, req3) AS max FROM invoices WHERE max(req1, req2, req3) <= now()
SELECT * from (select max(req1, req2, req3) AS max FROM invoices) WHERE max <= now()

Whichever seems clearer to you, I think postgres actually runs the two the
same way. Note, you should mark your function immutable so postgres knows it
can optimize the second case into the first.

Doesn't:

SELECT max(req1, req2, req3) AS ma1 FROM invoices HAVING ma1 <= now();

work?

Now I'm going to have to go see if I've got that wrong as well...

Obviously I got the wrong end of the stick but any way, I was wrong that column
aliases could be used for group by and having clauses. At least that's what
7.2.3 is telling me.

--
Nigel J. Andrews