On Tue, 2010-02-09 at 10:46 +1100, Andrew McNamara wrote:
> The problem is deeper than that - when query parameters use the binary
> option, the server has no way to decode the binary parameter without an
> appropriate type OID.

Postgres does not attempt to decode anything (text or binary format)
until it figures out what type it is.

> As you say, postgres will cast types depending on context, however this
> is stricter when binary parameters are used (because they only have one
> valid interpretation, whereas a text parameter may have several).

Type casts are a different matter; they are only done after the unknown
literals' types have been determined:

  create table n(i int);

  -- insert numeric literal, which is cast to int (assignment cast)
  insert into n values(5.0); -- succeeds

  -- insert unknown literal, which is inferred to be of type int
  insert into n values('5.0'); -- fails on integer type input function
  ERROR:  invalid input syntax for integer: "5.0"
  LINE 1: insert into n values('5.0');


Can you provide a concrete example in which the text versus binary
format changes the type inference behavior?

Regards,
	Jeff Davis

In response to

• Re: Confusion over Python drivers at 2010-02-08 23:46:54 from Andrew McNamara

Responses

• Re: Confusion over Python drivers at 2010-02-09 01:51:09 from Andrew McNamara

pgsql-hackers by date

Next:	From: Tom Lane	Date: 2010-02-09 00:35:43
	Subject: Re: Order of operations in lazy_vacuum_rel
Previous:	From: Jeff Davis	Date: 2010-02-08 23:50:53
	Subject: Re: Confusion over Python drivers