sqlobject-discuss — All things SQLObject-related

[SQLObject] Question about Session Error

[Cary F. <ckf...@ho...>]

Hello,
Another bloke posted a while ago as well and said he hadn't received an 
answer.

What is going on here?   Have you ever seen this error before? I think that 
there is some problem in getting the dirpath from the HTTP header, but don't 
know enough about this to figure it out.
If I can't get this to work I'll have to bag SQLObject, which would be a 
loss since it has so much potential.

This error occurs when I use the publisher handler to view a page, and in 
that function it tries to create a session object, which is on line 24.
Any ideas?

Mod_python error: "PythonHandler mod_python.publisher"

Traceback (most recent call last):

  File "C:\PYTHON23\Lib\site-packages\mod_python\apache.py", line 299, in 
HandlerDispatch
    result = object(req)

  File "C:\PYTHON23\Lib\site-packages\mod_python\publisher.py", line 136, in 
handler
    result = util.apply_fs_data(object, req.form, req=req)

  File "C:\PYTHON23\Lib\site-packages\mod_python\util.py", line 361, in 
apply_fs_data
    return object(**args)

  File "C:/Program Files/Apache Group/Apache2/htdocs\Login.py", line 24, in 
loadPage
    sess = Session(req, secret=SECRET_CODE)

  File "C:\PYTHON23\Lib\site-packages\mod_python\Session.py", line 386, in 
Session
    timeout=timeout, lock=lock)

  File "C:\PYTHON23\Lib\site-packages\mod_python\Session.py", line 358, in 
__init__
    timeout=timeout, lock=lock)

  File "C:\PYTHON23\Lib\site-packages\mod_python\Session.py", line 132, in 
__init__
    Cookie.add_cookie(self._req, self.make_cookie())

  File "C:\PYTHON23\Lib\site-packages\mod_python\Session.py", line 160, in 
make_cookie
    c.path = dirpath[len(docroot):]

TypeError: unsubscriptable object

Cary FitzHugh
Software Engineer

_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar – get it now! 
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/

Re: [SQLObject] Quesiton about multiple Joins

[Ivo v. d. W. <iv...@am...>]

On 15-jun-04, at 16:52, Cary Fitzhugh wrote:

> Hey all,
>
> I'm using SQLObject 0.5 with SQLite and had a problem with multiple 
> joins.
>
> The relevant code is:
>
> class Users(SQLObject):
>    _connection = conn
>    name = StringCol(default=None)
>    email = StringCol(default=None)
>    password = StringCol()
>    username = StringCol(alternateID=True)
>    lastLogin = DateTimeCol(default=None)
>    user_type = IntCol(default=None)
>    projects = MultipleJoin('Projects')
>
> AND:
>
> class Projects(SQLObject):
>    _connection = conn
>    state = IntCol()
>    number = IntCol(alternateID=True)
>    creator = ForeignKey('Users')
>    client = ForeignKey('Clients',notNone=True)
>
> As far as I could tell in the documentation to connect these two 
> tables, I just say:
> Projects.add(blah blah, creator=Users(1))
>
> and then
> for a in Users(1).projects:
>        print a
>
> should give me something, but it just says that there is no users_id 
> in the projects table.
>

Add joinColumn="creator_id' in your MultipleJoin definition.

Cheers

	Ivo

[SQLObject] Old data fetched by SQLObject in Webware.

[<mic...@ya...>]

Hi guys!

I have a problem integrating webware and sqlobject. I think the
question is the webware caching, but I don´t know well, look at the
code:

##############################
from Database import SomeTable
from SearchForm import SearchForm

class Search(SearchForm):
  def validateRequest(self):
    ...
    self.result=SomeTable.select(SomeTable.q.some_column=='bla')

  def writeContent(self):
    for i in self.result:
      self.writeln('some information...')
##############################

The SearchForm is a Form based on a Template, this one based on page...
On my template I defined a method "validateRequest", it verifies if
there is an error and append the error to a error list, if this error
list is true, I show the error... 

One time I need to check how many rows I get, and a proper time show a 
error massage in the error case, I need to generate the self.result in
the validateRequest method.

The problem is that the database changes (every time, outside the
webware) and the data shown in the Search are not changed. I have some
similar problem in another project, in this case, the database don´t
change every time but when I change it by using
SomeTable.get(x).set(...), some sectors of the page have the proper
data and some sections not.

What can I do, please? Thanks for any help :)

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br

[SQLObject] Quesiton about multiple Joins

[Cary F. <ckf...@ho...>]

Hey all,

I'm using SQLObject 0.5 with SQLite and had a problem with multiple joins.

The relevant code is:

class Users(SQLObject):
    _connection = conn
    name = StringCol(default=None)
    email = StringCol(default=None)
    password = StringCol()
    username = StringCol(alternateID=True)
    lastLogin = DateTimeCol(default=None)
    user_type = IntCol(default=None)
    projects = MultipleJoin('Projects')

AND:

class Projects(SQLObject):
    _connection = conn
    state = IntCol()
    number = IntCol(alternateID=True)
    creator = ForeignKey('Users')
    client = ForeignKey('Clients',notNone=True)

As far as I could tell in the documentation to connect these two tables, I 
just say:
Projects.add(blah blah, creator=Users(1))

and then
for a in Users(1).projects:
        print a

should give me something, but it just says that there is no users_id in the 
projects table.

I am guessing that there is some flag I am not setting or something I'm not 
doing that would make that flag get added.

Any suggestions?

_________________________________________________________________
Get fast, reliable Internet access with MSN 9 Dial-up – now 3 months FREE! 
http://join.msn.click-url.com/go/onm00200361ave/direct/01/

[SQLObject] Re: Transaction semantics do not appear to be working

[Ahmed M. A. <ahm...@wa...>]

Hi,
The connection object that you are using is your original connection,
the transaction create a new connection,so you must use that one.
The Transaction class uses internally this connection.It exposes the
same interface than the Connection class itself and uses a delegation
mechanism
to communicate with the database server.Here's how you can modify your code
in order to work with transaction:

<begin modified code>
trans = WKVTestTable._connection.transaction()
for i in xrange(1000):
 newrec = WKVTestTable(connection=trans,...) --> set the connection argument
to be the transaction itself
trans.commit()
</end modified code>

Ahmed MOHAMED ALI

"William Volkman" <wkv...@ne...> wrote in message
news:108...@ch......
> I've a small program that inserts a couple of thousand rows into
> a postgres database.  It appears that transaction handling is
> not working as I can see intermediate rows in the database
> before the program has committed them.  I believe that I'm following
> the correct sequence of commands:
>
> conn = connectionForURI("postgres://wkv@localhost/testdb")
> class WKVTestTable(SQLObject):
>     _connection = conn
>     _table = 'testtable'
>
>     id = IntCol(dbName='testid')
>     officeid = IntCol(dbName='officeid')
>     pointid = IntCol(dbName='pointid')
>     dest = StringCol(length=1, varchar=False, dbName='dest')
>     origin = StringCol(length=1, varchar=False, dbName='origin')
>     createdate = DateCol(dbName='createdate', default=func.CURRENT_DATE)
>     sourcedate = DateCol(dbName='sourcedate', default=None)
>
>
> trans = WKVTestTable._connection.transaction()
>
> for i in xrange(1000):
> newrec = WKVTestTable(...)
> trans.commit()
>
> I did not specify _cacheValues=False as that would seem to only
> apply to a multi access environment (perhaps I'm mistaken?)
>
> I've updated to revision 126 just now and still no joy.  Is
> autocommit being set on the connection?  Postgresql V7.3.4 on
> RH 9, psycopg-1.1.11-py2.2_RH9_wkv1
>
> Any suggestions?
>
> Thanks,
> William.
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by the new InstallShield X.
> From Windows to Linux, servers to mobile, InstallShield X is the
> one installation-authoring solution that does it all. Learn more and
> evaluate today! http://www.installshield.com/Dev2Dev/0504

[SQLObject] maxdb support

[Ahmed M. A. <ahm...@wa...>]

Hi all,
Sorry about this other message.I 'am making a new one because my previous
post is not appearing to be sent.
I added the maxdb support to sqlobject.This addon reside in the svn trunk.

[SQLObject] SQLObject findings

[Ivo v. d. W. <iv...@am...>]

Hi All,

I've been working with SQLObject for a while now and I love it. I have some 
remarks and questions after using it for a while though:

Views:

For what it's worth - I tried to create an SQLObject on top of a view (in 
PostgreSQL). This worked nicely for read operations, but updates don't work, 
partially because SQLObject assumes that Myview.id has a sequence myview_id_seq, 
while the original sequence is based on a different table on top of which the 
view is defined.

Nothing serious, just something I tried. Perhaps someone has any real purpose 
for this :)

SQL Inheritance:

I tried creating a tree-like structure in SQL with a base 'Node' table and 
derived tables that need to be linked together. SQLObject can't really handle 
this. Lots of (SQL) ways to work around this, so no real problem, but at that 
time, my schema would have been more readable (and it would have allowed me to 
play with SQL inheritance)

Inherticance (FAQ):

It's not quite clear to me if the section 'How Does Inheritance Work?' in the 
FAQ relates to SQL Inheritance or to Python inheritance. Simple (single) 
inheritance in python works fine for me, i.e.

class ParentRelation(SQLObject):
     parent = ForeignKey("Module")
     priority = IntCol(default=0)

class ModuleParentRelation(ParentRelation):
     child = ForeignKey("Module")

ModuleParentRelation.createTable()

This creates just a module_parent_relation table, which is what I would expect.

However, multiple inheritance (i.e. Mixins) doesn't work - the mixed in 
attributes just don't get generated in SQL, i.e.

class foo:
   f = StringCol()
class bar:
   b = StringCol()
class blah(SQLObject, foo, bar) # or (foo, bar, SQLObject)
   bl = StringCol()

Doesn't do the trick (come to think of it - should I derive foo and bar from 
SQLObject as well? I should give that a try)

SQLBuilder:

I'm using SQLBuilder to retrive childs of my nodes (which are ordered), 
something like:

     def Modules(self):
         return Module.select(AND(ModuleParentRelation.q.childID == Module.q.id,
                                  ModuleParentRelation.q.parentID == self.id),
                                  orderBy="module_parent_relation.priority")

Why can't orderBy use the q attribute of the class? I.e. 
orderBy=ModuleParentRelatio.q.priority? (Or is there another way?)

Now I need to lookup the generated tablename, and it looks less readable.

Name genration:

I noticed that all-caps class names result in a rather weird generated class 
name. I.e. 'FOO' will result in a class named 'fo_o'.

This is most likely completely according to the defined naming scheme, but I 
wonder if anyone really wants this.

All in all, a great tool, I enjoy it alot.

Cheers

	Ivo




-- 
Drs. I.R. van der Wijk                                      -=-
Korte Leidsedwarsstraat 12                     Amaze Internet Services V.O.F.
1017 RC Amsterdam, NL                                       -=-
T +31-20-4688336         F +31-20-4688337       Zope/Plone/Content Management
W http://www.amaze.nl    E in...@am...           Open Source Solutions
W http://vanderwijk.info E iv...@am...                 Consultancy
PGP http://vanderwijk.info/pgp

[SQLObject] Transaction semantics do not appear to be working

[William V. <wkv...@ne...>]

I've a small program that inserts a couple of thousand rows into
a postgres database.  It appears that transaction handling is
not working as I can see intermediate rows in the database
before the program has committed them.  I believe that I'm following
the correct sequence of commands:

conn = connectionForURI("postgres://wkv@localhost/testdb")
class WKVTestTable(SQLObject):
    _connection = conn
    _table = 'testtable'
 
    id = IntCol(dbName='testid')
    officeid = IntCol(dbName='officeid')
    pointid = IntCol(dbName='pointid')
    dest = StringCol(length=1, varchar=False, dbName='dest')
    origin = StringCol(length=1, varchar=False, dbName='origin')
    createdate = DateCol(dbName='createdate', default=func.CURRENT_DATE)
    sourcedate = DateCol(dbName='sourcedate', default=None)


trans = WKVTestTable._connection.transaction()

for i in xrange(1000):
	newrec = WKVTestTable(...)
trans.commit()

I did not specify _cacheValues=False as that would seem to only
apply to a multi access environment (perhaps I'm mistaken?)

I've updated to revision 126 just now and still no joy.  Is
autocommit being set on the connection?  Postgresql V7.3.4 on
RH 9, psycopg-1.1.11-py2.2_RH9_wkv1

Any suggestions?

Thanks,
William.

[SQLObject] SQLite and :memory:

[Daniel B. <da...@cs...>]

I'm playing with today's svn, and the db connection URIs (used with
connectForURI) for sqlite don't seem to handle the special
memory-resident target, which sqlite calls :memory:.

Normally, if you tell sqlite to use a file named :memory:, it will keep
the database in memory instead of using a file on disk. This worked fine
with the SQLiteConnection(file=':memory:') interface, but the current
URI parsing can't handle it (DBConnection._parseURI).

Some examples:

  URI:       sqlite:/:memory:
  _parseURI: user=None password=None host=None path=/:memory: args={}

  URI:       sqlite://:memory:
  _parseURI: user=None password=None host=:memory: path=/ args={}

  URI:       sqlite:///:memory:
  _parseURI: user=None password=None host=None path=/:memory: args={}

In none of these does the path come out as :memory:. Moreover, the first
and third examples should probably be considered correct, since those
URIs identify the file /:memory: , and not SQLite's pseudo-thing that
indicates the lack of a file.

But in the second URI, the parser interprets an identifer after
schema:// as the host, and expects the path to follow another '/' (which
isn't supplied, in this example).

Since hosts have no other meaning in sqlite URIs, maybe accepting a host
named :memory: would be an acceptable hack?

Here is a (line-broken) `svn diff` of my implementation of this idea. I
also removed the addition of an extra leading '/', which
DBConnection._parseURI already provides:

 danb


Index: sqlobject/sqlite/sqliteconnection.py
===================================================================
--- sqlobject/sqlite/sqliteconnection.py        (revision 126)
+++ sqlobject/sqlite/sqliteconnection.py        (working copy)
@@ -21,10 +21,14 @@
         DBAPI.__init__(self, **kw)
  
     def connectionFromURI(cls, uri):
+        magic_host = ':memory:'
         user, password, host, path, args = cls._parseURI(uri)
-        assert host is None, "SQLite can only be used locally (with a
URI like sqlite:///file or sql:/file, not %r)" % uri
+        assert host in [ None, magic_host ], "SQLite can only be used
locally (with a URI like sqlite:///file or sql:/file, not %r), or with
the magic host '%s' to specify in-memory use." % (uri, magic_host)
         assert user is None and password is None, "You may not provide
usernames or passwords for SQLite databases"
-        return cls(filename='/' + path, **args)
+        if host == magic_host:
+            assert path == '/', "You may not provide a path with the
magic host '%s'" % magic_host
+            path = host
+        return cls(filename=path, **args)
     connectionFromURI = classmethod(connectionFromURI)
  
     def _setAutoCommit(self, conn, auto):


-- 
Daniel Brown
www.utacm.org
www.cs.utexas.edu

[SQLObject] Re: multiple databases

[Justus P. <ju...@ry...>]

Here's a more complete (I think I got everything) and somewhat tested 
version of the same patch.  This adds connection keyword parameters to 
all of the class methods in SQLObject.

--
Justus

Re: [SQLObject] plpgsql integration.

[<mic...@ya...>]

Hi again guys!

I search through gmane and I can found a message can help me, so I
could do something, but I want to know if I do in the correct way, i
will show all the code:

#########################################
from sqlobject import *
conexao='postgres://localhost/teste'

class Person(SQLObject):
  _connection=conexao
  name=StringCol(length=100)
  areas=RelatedJoin('Area')
  def byArea(self, *elements):
    connection=self._connection
    in_clause=''
    for i in elements:
      in_clause+='%s, ' % i
    result=connection.queryAll('''\
select
  p.id
from person p
  inner join area_person x on (x.person_id=p.id)
  inner join area a on (x.area_id=a.id)
where 
  a.id in (%s)
group by
  p.id;
''' % in_clause[:-2])
    return IN(self.q.id, [x[0] for x in result])
  byArea=classmethod(byArea)

class Area(SQLObject):
  _connection=conexao
  name=StringCol(length=100)
  people=RelatedJoin('Person')
#########################################

so I can do:

[x.id for x in Person.select(Person.byArea(1, 2))]

I can return the selected object (but i prefer to run select manually,
so I can agregate another conditions) and I can receive the Area
instance, but for a test this code is ok I think...

Is there another way to run this query? Is this the best way to execute
queries, see that I can do, reater than execute a query, execute a
function...

thanks for the attention and sorry my poor english :)

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br

[SQLObject] plpgsql integration.

[<mic...@ya...>]

Hi guys!

First sorry about my 3 identical and consecutive messages, I don't know
why but I was not receiving messages and I think the problem is with my
email, but the sending was ok :)

I read a few days a message talking about random execute of querys on
the database.

I want to execute some complex join, or a plpgsql function and the
return I want to pass to the class (i want to do this in a class
method)...

I search on the history and didn't found. How can I do this?

Example:

class Person(SQLObject):
  _connection=conexao
  name=StringCol(length=100)
  areas=RelatedJoin('Area')
  def byArea(self, *elements):
    # do something with this query
    query='''\
select 
  p.*
from person p
  inner join area_person x on (x.person_id=p.id)
  inner join area a on (x.area_d=a.id)
where 
  a.id in %s 
group by 
  p.id, p.name
''' % tuple(elements))

class Area(SQLObject):
  _connection=conexao
  name=StringCol(length=100)
  people=RelatedJoin('Person')

I want to filter all the people in a determined area, I can't do it
with normal SQLObject aproach, so the only way I found is this way...

I could need to run a plpgsql query in place of the listed, so I wonder
how can I run this query :)

thanks for all help!

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br

[SQLObject] multiple databases

[Justus P. <ju...@ry...>]

I've been trying to use multiple databases simultaneously within 
SQLObject and it works pretty well thus far except for the creation and 
deletion of tables, which don't take a connection parameter.  Looking 
through the code I see that there is a comment "Should these have a 
connection argument?"

If I say pretty please can I get those methods to take a connection 
parameter? :-)

I've attached a completely untested patch that does this.

--
Justus

Re: [SQLObject] rough DB2 version of DBConnection.py

[Andy T. <an...@ha...>]

Peter Gebauer wrote:
>>is anybody even interested in using SQLObject with DB2? 
> 
> 
> We're upgrading our IBM DB2 8.1 database to PostgreSQL 7.4, but the old 
> system is still running DB2. It's unfortunate that we run DB2 at all (I'm 
> very unhappy with it) and one of the more painful things is the lack of good 
> DB2 Python support.
> It would be most welcome to have a nice Python interface to DB2 since we are 
> hacking PERL at the moment.
> 
> /Peter
> 
> 

There is a Python module for DB2 at http://sourceforge.net/projects/pydb2/

You might also want to look at mxODBC 
(http://www.egenix.com/files/python/mxODBC.html) it works with DB2 UDB 
but does require a commercial license if you are using it for anything 
other than personal use.

Regards,
Andy
-- 
--------------------------------------------------------------------------------
 From the desk of Andrew J Todd esq - http://www.halfcooked.com/

[SQLObject] SQLite and :memory:

[Lars H. <sql...@qu...>]

Hi all,

SQLite supports databases in memory (with :memory: as 'filename'). Is
it possible to use this feature with SQLObject?

Thanks,
Lars

Re: [SQLObject] patch for pgPySQL

[Jean-Philippe Barrette-L. <jpb...@sa...>]

On June 8, 2004 02:39 pm, Jean-Philippe Barrette-LaPierre wrote:
> I modified SQLObject to be able to use pgPySQL. here's the patch. If you
> don't like it, just tell me what's wrong and I'll change it to fullfill
> your needs.
sorry... forgot to attach the file.
-- 
Jean-Philippe Barrette-LaPierre
Maintener of curlpp

[SQLObject] patch for pgPySQL

[Jean-Philippe Barrette-L. <jpb...@sa...>]

I modified SQLObject to be able to use pgPySQL. here's the patch. If you don't 
like it, just tell me what's wrong and I'll change it to fullfill your needs.

-- 
Jean-Philippe Barrette-LaPierre
Maintener of curlpp

[SQLObject] Missing TimeCol ?

[William V. <wkv...@ne...>]

After not finding a tool to create the python classes from a Postgres
database I decided to write one (a DBA centric approach to development,
they define the what, the programmer builds interfaces).  I find that
the "time" data type is missing (The "DateCol" is there which should
be added to the Docs).

I also note that the 6.0 docs mention the consideration of multi field
primary keys which I'm hoping gets implemented.  As a data point the
schema I'm working with has 120 tables, only 6 meet the requirement
of an single integer primary key field. I've wrangled the DBA into
adding two more however he's very much against adding it to the
table that gets the most use.  So for now I'm hand wrapping some
tables, it certainly would be nice to use SQLObject instead.

Thanks,
William.

[SQLObject] Using RelatedJoin.

[<mic...@ya...>]

Hi guys!

I want some help with this join:

class Person(SQLObject):
  name=StringCol(length=100)
  specific_areas=RelatedJoin('SpecificArea')

class GeneralArea(SQLObject):
   name=StringCol(length=50)
  specific_areas=MultipleJoin('SpecificArea')

class SpecificArea(SQLObject):
  name=StringCol(length=50)
  general_area=ForeignKey('GeneralArea')
  people=RelatedJoin('Person')

I want to use some join like this:

select * from person p inner join person_specific_area x on
p.id=x.person_id inner join specific_area s on
s.id=x.specific_area_id
inner join gereral_area g on s.general_area_id=g.id;

I don´t try it, but I think the join it´s the above, I want to add
some
restrictions on it: 'where s.id=someid' (I want to filter people by
his
specific_areas).

I can use the IN() function of the SQLBuilder, but I can´t create the
join with the 2 tables, I don´t know what may I pass to
Person.select()...

thanks for all help and sorry about my poor english!

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br

[SQLObject] Using RelatedJoin.

[<mic...@ya...>]

Hi guys!

I want some help with this join:

class Person(SQLObject):
  name=StringCol(length=100)
  specific_areas=RelatedJoin('SpecificArea')

class GeneralArea(SQLObject):
  name=StringCol(length=50)
  specific_areas=MultipleJoin('SpecificArea')

class SpecificArea(SQLObject):
  name=StringCol(length=50)
  general_area=ForeignKey('GeneralArea')
  people=RelatedJoin('Person')

I want to use some join like this:

select * from person p inner join person_specific_area x on
p.id=x.person_id inner join specific_area s on
s.id=x.specific_area_id
inner join gereral_area g on s.general_area_id=g.id;

I don´t try it, but I think the join it´s the above, I want to add some
restrictions on it: 'where s.id=someid' (I want to filter people by his
specific_areas).

I can use the IN() function of the SQLBuilder, but I can´t create the
join with the 2 tables, I don´t know what may I pass to
Person.select()...

thanks for all help and sorry about my poor english! 

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br

Re: [SQLObject] Date utils

[Neil Blakey-M. <nb...@mi...>]

On Wed 2004-06-02 (15:15), Philippe Normand wrote:
> > a = sql.table.acct
> > if start_date:
> >     criteria.append(a.stamp_inserted >= start_date)
> > if end_date:
> >     criteria.append(a.stamp_inserted < end_date)
> > 
> 
> But it doesn't change the problem, too many objects are fetched from the db.
> Or am I missing something ?
> 
> I'd like to compare dates directly in select() and not in select() results.
> Following your code snippet, I'll try something like:
> 
> table.select((Table.q.stamp_inserted >= start_date) &
>              (Table.q.stamp_inserted < end_date))

Sorry, I tend to build up a criteria list, and go:

	table.select(AND(*criteria))

So yes, it will work for select queries just fine.  Just mind the bug if
seconds are at all important to you.

Neil
-- 
Neil Blakey-Milner
nb...@mi...
http://mithrandr.moria.org/

Re: [SQLObject] Date utils

[Philippe N. <sw...@fr...>]

On Wed, Jun 02, 2004 at 02:13:58PM +0200, Neil Blakey-Milner wrote:
> > 	
> > 
> > it isn't really efficient because it fetches all articles. Is there a clean
> > way to select SQLObjects given a specific date interval on DateTimeCol ?
> 
> I just compare them to a datetime, and that works just fine for me.
> Although, there's a bug where the datetime conversion uses '%s' instead
> of '%S' for seconds, which probably wouldn't affect your usage.
> 
> a = sql.table.acct
> if start_date:
>     criteria.append(a.stamp_inserted >= start_date)
> if end_date:
>     criteria.append(a.stamp_inserted < end_date)
> 

But it doesn't change the problem, too many objects are fetched from the db.
Or am I missing something ?

I'd like to compare dates directly in select() and not in select() results.
Following your code snippet, I'll try something like:

table.select((Table.q.stamp_inserted >= start_date) &
             (Table.q.stamp_inserted < end_date))

Cheers,

Philippe

Re: [SQLObject] Date utils

[Neil Blakey-M. <nb...@mi...>]

On Wed 2004-06-02 (14:05), Philippe Normand wrote:
> I see one durty solution which would like this (untested) :
> 
> calendar = {} # keys: day numbers of the month
> for article in Article.select(orderBy=Article.q.date):
>     if article.getYear() != 2004 or article.getMonth() != 5:
>        continue
>     try:
>         calendar[article.getDay()] += 1
>     except KeyError:
>         calendar[article.getDay()] = 1
> 	
> 
> it isn't really efficient because it fetches all articles. Is there a clean
> way to select SQLObjects given a specific date interval on DateTimeCol ?

I just compare them to a datetime, and that works just fine for me.
Although, there's a bug where the datetime conversion uses '%s' instead
of '%S' for seconds, which probably wouldn't affect your usage.

a = sql.table.acct
if start_date:
    criteria.append(a.stamp_inserted >= start_date)
if end_date:
    criteria.append(a.stamp_inserted < end_date)

Neil
-- 
Neil Blakey-Milner
nb...@mi...
http://mithrandr.moria.org/

[SQLObject] Date utils

[Philippe N. <sw...@fr...>]

Hi,

I need to sort and count SQLObjects per date. An example might really show
what I want :)

class Article(SQLObject):
    date = DateTimeCol()
    ...
    
    
I have many many Articles, and I'd like to build a Calendar showing how many
articles were published by month and/or year. For instance:

10 articles on 05/22
3 articles on 05/14

I see one durty solution which would like this (untested) :

calendar = {} # keys: day numbers of the month
for article in Article.select(orderBy=Article.q.date):
    if article.getYear() != 2004 or article.getMonth() != 5:
       continue
    try:
        calendar[article.getDay()] += 1
    except KeyError:
        calendar[article.getDay()] = 1
	

it isn't really efficient because it fetches all articles. Is there a clean
way to select SQLObjects given a specific date interval on DateTimeCol ?

Philippe

[SQLObject] Using RelatedJoin.

[<mic...@ya...>]

Hi guys!

I want some help with this join:

class Person(SQLObject):
  name=StringCol(length=100)
  specific_areas=RelatedJoin('SpecificArea')

class GeneralArea(SQLObject):
  name=StringCol(length=50)
  specific_areas=MultipleJoin('SpecificArea')

class SpecificArea(SQLObject):
  name=StringCol(length=50)
  general_area=ForeignKey('GeneralArea')
  people=RelatedJoin('Person')

I want to use some join like this:

select * from person p inner join person_specific_area x on
p.id=x.person_id inner join specific_area s on s.id=x.specific_area_id
inner join gereral_area g on s.general_area_id=g.id;

I don´t try it, but I think the join it´s the above, I want to add some
restrictions on it: 'where s.id=someid' (I want to filter people by his
specific_areas).

I can use the IN() function of the SQLBuilder, but I can´t create the
join with the 2 tables, I don´t know what may I pass to
Person.select()...

thanks for all help and sorry about my poor english!

=====
--
Michel Thadeu Sabchuk
Curitiba/PR

______________________________________________________________________

Participe da pesquisa global sobre o Yahoo! Mail: 
http://br.surveys.yahoo.com/global_mail_survey_br