[Zope] - ANNOUNCE: DCOracle release

[Jim Fulton]

Kevin Littlejohn wrote:
> 
> >>> Anthony Baxter wrote
> > I'd guess that one of the results is larger than can fit in an integer,
> > maybe?
> 
> Actually, it looks like it was in dealing with the input parameters, rather
> than the output ones - I found in ociCurs.py, around line 124, a piece of
> code that turns a number into an integer if it has no 'scale'.  On commenting
> that out, it appears to work.  If someone with a touch more knowledge than
> I can comment on the ramifications of that, it'd be appreciated ;)

The ramification is that all numbers are treated as floats.

I think that oracledb treats all numbers as floats.

IMO, if I have:

  create table spam (val integer);

  insert into spam values (1);

  select * from spam;

I should get integers back in Python.  Unfortunately, 
Oracle says that val is a number with precision 38
and scale 0.  For most applications it should be
OK to use integers, but sometimes it's not and it's hard
(impossible) to determine when integers should be used 
without fetching the data first. :(

I'll have to look at this.  I should probably
use Oracles NUMBER or VARNUM data types to fetch
data and then convert to Python integers or longs, 
based on the values.

If anyone wants to help out with this, I'd be happy
to exchange information.

(snip)

> Oh, one other thing I wanted to double-check:  In db.py (back in ZOracleDA),
> in the 'columns' method, there's the following code:
> 
>     def columns(self, table_name):
>         c=self.cursor
>         try: r=c.execute('select * from %s' % table_name)
>         except: return ()
>         desc=c.description
> 
> That looks like tempting fate to me...  Is the 'select * from ' really
> not going to do bad things?  (I'm presuming it doesn't actually trigger
> the db into _doing_ a 'select * from' unless it calls fetchone()) 

This is my understanding also.

> Is there a less, um, dangerous-looking way of getting the same result?

It doesn't matter if it's dangerous looking if it's not actually
dangerous. :) 

I use this approach based on a recommendation made for ODBC.
The people who wrote Intersolve wrote a document on optimizing
ODBC calls and recommended doing selects rather than using ODBC
cataloging functions.  This approach also has the bonus
that it fails if you don't have access to the table, which
is a good thing in some cases. :)

Alternatively, an Oracle data dictionary table could be queried.
I assume that this is six one half-a-dozen the other.

> I only ask because select * from most of the tables here would go away
> for a _very_ long time...

Have you tried this?  Have you tries calling columns
for one of these tables?

Jim

--
Jim Fulton           mailto:jim@digicool.com
Technical Director   (888) 344-4332              Python Powered!
Digital Creations    http://www.digicool.com     http://www.python.org

Under US Code Title 47, Sec.227(b)(1)(C), Sec.227(a)(2)(B) This email
address may not be added to any commercial mail list with out my
permission.  Violation of my privacy with advertising or SPAM will
result in a suit for a MINIMUM of $500 damages/incident, $1500 for
repeats.