Re: Getting adaped output from unicode text, without a connection

On Sun, Jul 21, 2013 at 7:56 PM, Hefferon, James S. <jhefferon(at)smcvt(dot)edu> wrote:
> Hello,
>
> I need to write some static .sql files (just as an explanation, it is startup code for Django). Some of the strings are non-ascii. I'm stumped about how to get utf-8 instead of latin-1 in the file.
>
> I'm generating the .sql as strings, without a connection. I found a prior message on this list that suggested what to do (1) and came up with this routine.
>
> def todb(s):
> psycopg2.extensions.register_type(psycopg2.extensions.UNICODE)
> psycopg2.extensions.register_type(psycopg2.extensions.DECIMAL)
> psycopg2.extensions.register_type(psycopg2.extensions.FLOAT)
> adapted = psycopg2.extensions.adapt(s)
> return adapted.getquoted()
>
> But it encodes in latin-1. I understand from the docs that I need to get a connection and set the .encoding attribute, but I can't see how to get a connection instance without a db handy. For instance,
>
>>>> conn = psycopg2.connect(database = None)
>
> fails.
>
> I expect that I am asking a dopey question (I'm not very sure what I am doing) but I'd appreciate a pointer even if it makes me feel dumb.

No, it's not a dumb question. But I'm afraid you cannot create a
psycopg connection object without a real database connection.

If you had a working connection you could adapt the string with:

conn.set_client_encoding('utf8')
adapted = psycopg2.extensions.adapt(s)
adapted.prepare(conn)
return adapted.getquoted()

you don't need the register_type calls above: they only control how
data from the database is converted to Python.

-- Daniele