sqlobject-discuss — All things SQLObject-related

[SQLObject] SQLObject, _perConnection & PicklingError

[Markos K. <mk...@gm...>]

So, I had a class that had an internal _connection. When I tried to pickle it, I got errors, which from the looks of main were related to whether I had a per instance connection or went through sqlhub. So I rewrote the class & code as follows:

connection = connectionForURI('mysql://%s:%s@%s:3306/%s?debug=1&use_unicode=1&charset=utf8' % (dbuser,dbpass,dbhost,dbname))
sqlhub.processConnection = connection

class Title(SQLObject):
    class sqlmeta:
        fromDatabase = True
    books = MultipleJoin('Book')
    author = RelatedJoin('Author', intermediateTable='author_title',createRelatedTable=True)
    categorys = MultipleJoin('Category')
    kind = ForeignKey('Kind')
    listTheseKeys=('kind')

but I'm still getting the same PicklingError. What can I be missing? I've seen a few recent posts about fixing this, but is there a workaround?

Thank much
--Marko

 /Users/mkapes/Downloads/infoshopkeeper/AmazonAuthorTitle.py in <module>()
     40 except KeyError:
     41     ourTitles=list(Title.select(RLIKE(Title.q.isbn, '^[0-9]{9}[0-9xX]{1}$') & (Title.q.kindID==1)))
---> 42     shelf['ourTitles']=ourTitles
     43 
     44 

/opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/shelve.pyc in __setitem__(self, key, value)
    130         f = StringIO()
    131         p = Pickler(f, self._protocol)
--> 132         p.dump(value)
    133         self.dict[key] = f.getvalue()
    134 

/opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/SQLObject-0.13dev_r4106-py2.6.egg/sqlobject/main.pyc in __getstate__(self)
   1679         if self.sqlmeta._perConnection:
   1680             from pickle import PicklingError
-> 1681             raise PicklingError('Cannot pickle an SQLObject instance that has a per-instance connection')
   1682         d = self.__dict__.copy()
   1683         del d['sqlmeta']

PicklingError: Cannot pickle an SQLObject instance that has a per-instance connection
> /opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/SQLObject-0.13dev_r4106-py2.6.egg/sqlobject/main.py(1681)__getstate__()
   1680             from pickle import PicklingError
-> 1681             raise PicklingError('Cannot pickle an SQLObject instance that has a per-instance connection')
   1682         d = self.__dict__.copy()

Re: [SQLObject] Python 3 version

[Oleg B. <ph...@ph...>]

On Sat, Feb 20, 2010 at 12:03:03PM +0200, Neil Muller wrote:
> On 19 February 2010 17:14, Oleg Broytman <ph...@ph...> wrote:
> >  I haven't started to move to Py3k myself. The biggest problem is that I
> > don't see DB AP drivers for Py3k
> 
> I have no idea about the status of MySQLdb, but psycopg did have
> several patches for python3 support applied a while ago. I believe
> it's current state is "mostly works with python 3".

   Thank you. "Mostly works" is a good state to start experimenting, but
not as good to start using it in real programs; seems I will not port my
programs to Python3 in the nearest future.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] Python 3 version

[Neil M. <drn...@gm...>]

On 19 February 2010 17:14, Oleg Broytman <ph...@ph...> wrote:
>   I haven't started to move to Py3k myself. The biggest problem is that I
> don't see DB AP drivers for Py3k (SQLObject uses DB API drivers to connect
> to backends and to pass SQL queries). I think sqlite3 driver is ok, but
> where can I download psycopg or MySQLdb for Python3 (these are the three
> major supported backend)? Are current versions of these drivers support
> Python3 out of the box? with some patches?

I have no idea about the status of MySQLdb, but psycopg did have
several patches for python3 support applied a while ago. I believe
it's current state is "mostly works with python 3".

-- 
Neil Muller
drn...@gm...

I've got a gmail account. Why haven't I become cool?

Re: [SQLObject] Python 3 version

[Oleg B. <ph...@ph...>]

Hello.

On Fri, Feb 19, 2010 at 03:00:18PM +0000, Alan Harris-Reid wrote:
> Can anyone tell me if a Python 3 version of SQLObject is available yet?  
> If not, does anyone know of any plans to port it?

   It seems I'm the only active developer left though there are many users;
among them there are people who send bug reports and even patches.
   I haven't started to move to Py3k myself. The biggest problem is that I
don't see DB AP drivers for Py3k (SQLObject uses DB API drivers to connect
to backends and to pass SQL queries). I think sqlite3 driver is ok, but
where can I download psycopg or MySQLdb for Python3 (these are the three
major supported backend)? Are current versions of these drivers support
Python3 out of the box? with some patches?
   If there were Python3-compatible drivers (sqlite3, at least) you can
test if SQLObject works with Py3k and report problems or even better send
patches. If the patches are compatible with Python 2 I can just include
them in the code. If they are not I will create a separate branch for
Python3 (and install Python3 myself to run tests).

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] Python 3 version

[Alan Harris-R. <aha...@go...>]

Hi,

Can anyone tell me if a Python 3 version of SQLObject is available yet?  
If not, does anyone know of any plans to port it?

Thanks,
Alan

Re: [SQLObject] SQLObject connection class for Rdbhost

[Oleg B. <ph...@ph...>]

Hello! I started to look at the patch. Almost everything is ok. Just one
question: why 'utf8' is added to kw? are you sure you always want it there?
in this spelling - 'utf8'? looks more like mysql...
   If you really want it there you don't need to test for it in kw at the
end of __init__.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] Commit on main connection

[Oleg B. <ph...@ph...>]

On Fri, Feb 12, 2010 at 08:57:06AM -1000, John Bresnahan wrote:
> I would like to disable autocommit on my underlying database connection.  What method do I use to 
> force a commit on the connection?  I realize that a transaction object has a commit() method, but 
> how do I do a similar thing with the connection i acquired via:
> 
> sqlhub.processConnection = connectionForURI(config.dburl)

   The object created by connectionForURI() is not a DB API connection but
an instance of SQLObject's class DBConnection; it's a high-level object
that remembers connection parameters; it opens and closes real DB
connections when they are needed, it maintains a pool of opened
connections (for multithreaded applications, e.g.). But it cannot be
committed because it's not a real connection (what low-level connection
from the pool are to be committed?)
   Use transactions.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

[SQLObject] Commit on main connection

[John B. <bre...@mc...>]

Hello,

I would like to disable autocommit on my underlying database connection.  What method do I use to 
force a commit on the connection?  I realize that a transaction object has a commit() method, but 
how do I do a similar thing with the connection i acquired via:

sqlhub.processConnection = connectionForURI(config.dburl)

Thanks,

John

Re: [SQLObject] InheritableSQLObject: Data corruption when child record create fails

[Oleg B. <ph...@ph...>]

Hi! I applied the patch with an additional test - that the connection is not
a transaction and is in autocommit mode. A transaction or non-autocommit
connection must be rolled back after a exception but the place for the
rollback is application-dependent so I didn't include a rollback in the
code. Inside a transaction it's also usually impossible to remove the
parent row(s) after an error - the only thing that's allowed is a rollback.
   Committed in the revisions 4101-4103 (branches 0.11, 0.12 ad the trunk).
Thank you!

On Mon, Dec 07, 2009 at 06:41:54PM +0200, Tom Coetser wrote:
> I am doing some tests with InheritableSQLobject and have run into a problem 
> which I hope to get some feedback on.
> 
> The problem is that when creating a new object based on an 
> InheritableSQLObject class, and the creation of the child object fails after 
> a new parent was created, the parent record will remain in the database 
> without the relevant child.
> 
> Here is a short example:
> 
> ~~~~~~~~~~ file: inheritTest3.py ~~~~~~~~~~~~
> from sqlobject import *
> from sqlobject.inheritance import InheritableSQLObject
> 
> __connection__ = "sqlite:/:memory:"
> 
> class Vehicle(InheritableSQLObject):
>     vid = StringCol(notNone=True, default=None, unique=True)
>     wheels = IntCol()
> 
> class Bicycle(Vehicle):
>     seats = IntCol(notNone=True, default=None)
> 
> class Car(Vehicle):
>     doors = IntCol(notNone=True, default=None)
> 
> def createTables():
>     Vehicle.createTable()
>     Bicycle.createTable()
>     Car.createTable()
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> Here is a sample session showing the problem when trying to create a bicycle 
> but only supplying the args required for the Vehicle part of the object:
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In [1]: from inheritTest3 import *
> 
> In [2]: createTables(True)
> 
> In [3]: b = Bicycle(vid="id1", wheels=4)
> ERROR: An unexpected error occurred while tokenizing input
> The following traceback may be corrupted or invalid
> The error message is: ('EOF in multi-line statement', (22, 0))
> 
> ---------------------------------------------------------------------------
> IntegrityError                            Traceback (most recent call last)
> 
> /home/tomc/mobilereporter.repo/main/app/<ipython console> in <module>()
> 
> /usr/lib/python2.5/site-packages/sqlobject/main.pyc in __init__(self, **kw)
>    1221                 id = None
>    1222
> -> 1223             self._create(id, **kw)
>    1224
> 
> .... snipping lots of traceback parts ............
> 
> /usr/lib/python2.5/site-packages/sqlobject/sqlite/sqliteconnection.pyc in 
> _executeRetry(self, conn, cursor, query)
>     187                 raise DuplicateEntryError(msg)
>     188             else:
> --> 189                 raise IntegrityError(msg)
>     190         except self.module.InternalError, e:
>     191             raise InternalError(ErrorMessage(e))
> 
> IntegrityError: bicycle.seats may not be NULL
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> If you now supply all required args for the bicycle, using the same field 
> values for the vid and wheels as before, a duplicate entry error is raised:
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> In [4]: b = Bicycle(vid="id1", wheels=4, seats=2)
> ERROR: An unexpected error occurred while tokenizing input
> The following traceback may be corrupted or invalid
> The error message is: ('EOF in multi-line statement', (22, 0))
> 
> ---------------------------------------------------------------------------
> DuplicateEntryError                       Traceback (most recent call last)
> 
> /home/tomc/mobilereporter.repo/main/app/<ipython console> in <module>()
> 
> /usr/lib/python2.5/site-packages/sqlobject/main.pyc in __init__(self, **kw)
>    1221                 id = None
>    1222
> -> 1223             self._create(id, **kw)
>    1224
>    1225             for func in post_funcs:
> 
> .... snipping lots of traceback parts ............
> 
> /usr/lib/python2.5/site-packages/sqlobject/sqlite/sqliteconnection.pyc in 
> _executeRetry(self, conn, cursor, query)
>     185             msg = ErrorMessage(e)
>     186             if msg.startswith('column') and msg.endswith('not 
> unique'):
> --> 187                 raise DuplicateEntryError(msg)
>     188             else:
>     189                 raise IntegrityError(msg)
> 
> DuplicateEntryError: column vid is not unique
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> The duplicate entry is because of the parent vehicle record was created when 
> the first bicycle creation failed: 
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In [5]: q = "SELECT * from vehicle"
> 
> In [6]: Vehicle._connection.queryAll(q)
> Out[6]: [(1, 'id1', 4, 'Bicycle')]
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> So, the problem is that if a new child object is created, but fails after the 
> parent has been created, the parent is left in tact in the database.
> 
> I'm attaching a patch as a possible fix for this, but I am not sure under 
> which other circumstances this patch will cause other problems. Maybe a flag 
> is needed inside the block that creates the child (parentCreatedHere=True), 
> and only if this flag is set, should the parent be deleted in exception if 
> the child creation failed?
> 
> I would appreciate if someone with more insight into this can look at the 
> patch an modify where neccessary.
> 
> BTW: This patch is against SQLObject 0.11.0.
> 
> 
> 
> Cheers,
>  Tom

> --- inheritance/__init__.py.orig	2009-12-07 14:47:31.000000000 +0000
> +++ inheritance/__init__.py	2009-12-07 16:35:39.000000000 +0000
> @@ -384,7 +384,19 @@
>  
>              id = self._parent.id
>  
> -        super(InheritableSQLObject, self)._create(id, **kw)
> +        # TC: Create this record and catch all exceptions in order to destroy
> +        # TC: the parent if the child can not be created.
> +        try:
> +            super(InheritableSQLObject, self)._create(id, **kw)
> +        except:
> +            # TC: If we are the child, destroy the parent
> +            if self.sqlmeta.parentClass:
> +                self._parent.destroySelf()
> +                #TC: Do we need to do this??
> +                self._parent = None
> +            # TC: Reraise the original exception
> +            raise
> +
>  
>      def _findAlternateID(cls, name, dbName, value, connection=None):
>          result = list(cls.selectBy(connection, **{name: value}))

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] TimeValidator from_python method missing?

[Oleg B. <ph...@ph...>]

On Mon, Feb 08, 2010 at 02:11:00PM +0200, Tom Coetser wrote:
> +    from_python = to_python
> +

   Thank you. I will test it.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

[SQLObject] TimeValidator from_python method missing?

[Tom C. <su...@ic...>]

Hi Oleg,

I'm just wondering if there is a reason why the TimeValidator class in Col.py 
does not also have the from_python method defined to be:

from_python = to_python

as is done for in the DateValidator class?

With this method defined in the DateValidator for example, one can set the 
date for a DateCol field using a string like "2010-01-31", without having to 
first instantiate a datetime.date() instance. The column validators chain 
would take care of converting the string to the correct type for the DateCol, 
depending on the column's dateFormat specification, of course.

But, the TimeValidator does not have the from_python method set like for the 
DateValidator, which means that one can not conveniently use string of the 
format "17:59:20" to set the value for a TimeCol(), and must first create and 
instance of datetime.time() in order to set the TimeCol() value.

Since SOTimeCol already have a timeFormat attribute which seems to be used in 
exactly the same way as the dateFormat attr for a DateCol, it seems that 
setting the from_python method on TimeValidator may just have been an 
oversight, or else there is another reason I am not aware of.

If it is an oversight, the small patch attached would allow strings to be used 
to set TimeCol values just as they are now conveniently usable for setting 
DateCol values.

IWBN if you can have a quick look and see if this can be added to future 
releases.



Cheers,
 Tom

Re: [SQLObject] how to force "select" method to retrieve just requested columns

[Petr J. <pet...@tp...>]

>
>   There is the way: .select() (and sqlbuilder.Select) accept a boolean
> parameters lazyColumns that does what you want! With lazyColumns=True
> .select() only SELECTs IDs, and then attributes will be loaded one by one
> or all at once depending on the cacheValues.
>

Thank a lot for reply. I have made a view in the database and I do not
care about number of columns (data cells) returned in one select
request any more.
The number of returned columns is "regulated" by the view SQL machinery.

Re: [SQLObject] manually created query

[Petr J. <pet...@tp...>]

>   You can execute the query but you cannot get the results as a list of
> SQLObject instances because there are no declarations of columns with sum()
> and other expressions. The result will be a list of tuples.
>

Yes, I know, but sometimes it is OK to get not-object  data (some
nasty complicated SQL queries for example)

Thanks Oleg, thanks Sean

Re: [SQLObject] manually created query

[Oleg B. <ph...@ph...>]

On Sun, Feb 07, 2010 at 09:06:42PM +0100, Petr Jake?? wrote:
> I would like to query the database something like this:
> 
> select sum(quantity) as qty, product_name, sum(quantity) * price *
> (1+tax_rate/100)
>         FROM `prestashop`.`ps_view_new_order_detail`
>         where id_order = 1
>         group by product_name
> 
> Is there a way to drop this "handcrafted" query to the SQLObject and
> get the result?

   You can execute the query but you cannot get the results as a list of
SQLObject instances because there are no declarations of columns with sum()
and other expressions. The result will be a list of tuples.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] manually created query

[Sean D. <hal...@gm...>]

> Is there a way to drop this "handcrafted" query to the SQLObject and
> get the result?
>


I think you can use Prestashop._connection.queryAll("SELECT ...")

~Sean

[SQLObject] manually created query

[Petr J. <pet...@tp...>]

Hi, I have feeling it has been discussed already in the past, but I
can not find the answer in the list archive.

I would like to query the database something like this:

select sum(quantity) as qty, product_name, sum(quantity) * price *
(1+tax_rate/100)
        FROM `prestashop`.`ps_view_new_order_detail`
        where id_order = 1
        group by product_name

I think it is not possible to create similar kind of queries through
the SQLObject.
(http://www.sqlobject.org/FAQ.html#what-about-group-by-union-etc)

Is there a way to drop this "handcrafted" query to the SQLObject and
get the result?

Regards

Petr

Re: [SQLObject] how to set dbEncoding for the whole table/database

[Petr J. <pet...@tp...>]

>
>   fromDatabase machinery for MySQL takes into account parameters from DB
> URI - 'use_unicode' and 'charset'. If use_unicode is set to True (1)
> columnsFromSchema() replaces all StringCol by UnicodeCol with charset as
> its dbEncoding.
>

Thanks. This works great.

Re: [SQLObject] how to set dbEncoding for the whole table/database

[Oleg B. <ph...@ph...>]

On Thu, Feb 04, 2010 at 10:15:27PM +0100, Petr Jake?? wrote:
> >   There is no currently. What do you want? To define an encoding for all
> > UnicodeCol'umns at once?
> >
> Yes, that was the main idea. When I am working with existing database, I do
> not declare columns manually (I am using fromDatabase = True).

   fromDatabase machinery for MySQL takes into account parameters from DB
URI - 'use_unicode' and 'charset'. If use_unicode is set to True (1)
columnsFromSchema() replaces all StringCol by UnicodeCol with charset as
its dbEncoding.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] how to force "select" method to retrieve just requested columns

[Oleg B. <ph...@ph...>]

On Sat, Feb 06, 2010 at 04:54:07PM +0100, Petr Jake?? wrote:
> MyTable.select() method results to the
> SELECT my_table.* FROM my_table (or ALL column names are named explicitly)
> 
> Is there a way to retrieve just some *columns*, not all of them.

   There is the way: .select() (and sqlbuilder.Select) accept a boolean
parameters lazyColumns that does what you want! With lazyColumns=True
.select() only SELECTs IDs, and then attributes will be loaded one by one
or all at once depending on the cacheValues.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] how to get id (primary key) from database

[Oleg B. <ph...@ph...>]

On Sat, Feb 06, 2010 at 12:20:10AM +0100, Petr Jake?? wrote:
> >   Yes, and idMessage/id_message must not be there. It is there because of
> > a bug. So help me to chase the bug.
> >
> > SHOW COLUMNS FROM ps_message_readed
> >
> >   (just run the query)
> 
> I see, I was trying to get it from within the Python.
> 
> Result attached.

   Thank you. The results look good. Method columnsFromSchema() should
recognize the id column 'id_message'... but it doesn't. Strange.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] how to force "select" method to retrieve just requested columns

[Oleg B. <ph...@ph...>]

On Sat, Feb 06, 2010 at 04:54:07PM +0100, Petr Jake?? wrote:
> MyTable.select() method results to the
> SELECT my_table.* FROM my_table (or ALL column names are named explicitly)
> 
> Is there a way to retrieve just some rows, not all of them.

   Some columns, you mean? No, SQLObject instance can only select all of
the column or none of them. Perhaps "none of them" is what you whant. Set

class sqlmeta:
    cacheValues = False

   Now every attribute access executes a separate SELECT query just to
select this column from the table. If you read only a few attributes there
will be only a few SELECTs.
   But I doubt it is faster or takes less memory than one wide SELECT. I'd
very much like to recommend to experiment and find out what way is really
faster.

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

Re: [SQLObject] Manage objects in different databases under the same program instance

[Oleg B. <ph...@ph...>]

On Sat, Feb 06, 2010 at 10:43:43AM -0300, Juan Manuel wrote:
> Can the .select() or .selectBy() methods take the connection to use as a
> parameter (even if I pass it a normal connection object, and not one
> obtained from a transaction? If that's true then all is well :)

   Yes. I have already said this - all SQLObject methods accept
'connection' parameter that must be a connection or a transaction object.
Default for the parameter is None meaning "obtain the connection from the
corresponding SQLObject class".

Oleg.
-- 
     Oleg Broytman            http://phd.pp.ru/            ph...@ph...
           Programmers don't die, they just GOSUB without RETURN.

[SQLObject] how to force "select" method to retrieve just requested columns

[Petr J. <pet...@tp...>]

MyTable.select() method results to the
SELECT my_table.* FROM my_table (or ALL column names are named explicitly)

Is there a way to retrieve just some rows, not all of them.

something what results to to the query like this:

SELECT my_table.columnA, my_table.columnB FROM my_table
(suppose there are also columns my_table.columnC, my_table.columnD ...
in the table).

Sometimes the tables I am working with are very wide (more than 100
columns) an I am connected them over the internet, so forcing the
select method to retrieve only requested rows will be great.

Best Regards

Petr

Re: [SQLObject] Manage objects in different databases under the same program instance

[Juan M. <vic...@gm...>]

On Sat, Feb 6, 2010 at 9:49 AM, Petr Jakeš <pet...@tp...> wrote:

>
>
> hmm... what about:
>
> http://www.sqlobject.org/SQLObject.html#sqlobject-class
> http://www.sqlobject.org/SQLObject.html#transactions
> http://www.sqlobject.org/SQLObject.html#dynamic-classes
>
> To be honest, I do not have experiences with this, but it looks the
> way to go ...
>

Can the .select() or .selectBy() methods take the connection to use as a
parameter (even if I pass it a normal connection object, and not one
obtained from a transaction? If that's true then all is well :)

Thanks!

Re: [SQLObject] Manage objects in different databases under the same program instance

[Petr J. <pet...@tp...>]

>
> As an example: currently when you scan a directory, there are two treeviews
> which are populated: the left one with the entire directory structure (only
> dirs), and the right one with the contents of the currently selected dir
> (much like windows explorer). If you scan another dir, the treeviews are
> wiped and all references to the previous scan are wiped too. My aim is to,
> instead of wiping all that, keep it and open a new tab (with another two
> treeviews), show the new catalog in that new tab, and be able to switch
> between tabs and vieweing both catalogs at the same time.T
>

hmm... what about:

http://www.sqlobject.org/SQLObject.html#sqlobject-class
http://www.sqlobject.org/SQLObject.html#transactions
http://www.sqlobject.org/SQLObject.html#dynamic-classes

To be honest, I do not have experiences with this, but it looks the
way to go ...