Re: Wierd "ProgrammingError" with "E" string prefix

On Wed, Mar 04, 2009 at 10:31:18AM -0800, Tim Roberts wrote:

> >> What version of Python are you running?  I've never seen an "E" prefix
> >> for strings in any context.
> >>     
> > It's not Python, it is PostgreSQL, and it's *inside* the
> > Python string that get's sent as the query.
> >   
> 
> I don't understand what you are saying.  Hartmut said he did this:
> 
>     curs.mogrify("SELECT %(foo)s AS foo", {'foo':'bar'} )
> and got this result
>     SELECT E'bar' AS foo
> 
> How is Postgres involved in that?  Isn't that entirely inside psycopg?

Well, the entirety of >>>SELECT E'bar' AS foo<<< is the
Python string (== the query) that is going to be sent to
PostgreSQL (if doing curs.execute(), that is).

Nonetheless, the E'bar' *inside* it is going to be
interpreted by PostgreSQL as a "string", in particular one
with C-like escapes:

	http://www.postgresql.org/docs/8.3/static/sql-syntax-lexical.html#SQL-SYNTAX-CONSTANTS

That's why psycopg2 generates E''s.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346