[sqlalchemy] Re: remote nondirect access to DB
Client application doesn't mean web application and SqlAlchemy would be a too big hammer for SQL console (in my opinion). Well it will be an information system written with wxPython running on multiple computers. I don't want users to see information for direct access to the database. I was trying to expose DBAPI (connections, cursors) via Pyro but it is falling. Sometimes with segfault. David Michael Bayer píše v Čt 07. 12. 2006 v 15:36 -0800: um, web interface ? SQL console ? it would help to know what kind of client youre talking about. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] guessing sql joins from object level
Hi, I am trying to translate to sqlalchemy my queries having on object level clauses like the a.b.c.d == some_value where a is instance of class A, b - instance of class B and... How such queries can be expressed in sqlalchemy code (supposing that every class is mapped to its own table)? Are there some possibility SA to automagically guess the needed joins when generating appropriate sql select? Thanks in advance StefanB --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ForeignKey and onupdate/ondelete
Michael Bayer ha scritto: use ForeignKeyConstraint. Yes, I have used it. But why ForeignKey does not allow this? Regards Manlio Perillo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: SA 0.3.x: performance issues = proposal
As attachement, here's the patch (against rev 2132). It's local to sqlalchemy/orm/, so: cd /sqlalchemy/orm patch -p0 attribute_cache.patch About the patch itself: 1. the cache should be a WeakKeyDictionary OK, done. Performances are still OK, differences between built-in dict are negligible. 2. the cache needs to be cleared for a particular cache (class ?) If I understand right, while calling register_attribute(), the cache will be cleared for the given class, if this one has been cached. Is there anywhere else the cache needs to cleared ? 3. the cache should *probably* be at the module level, and not within the AttributeManager itself; SA uses only a single AttributeManager per application so it probably doesnt matter I let the cache within the AttributeManager... Things can be moved at the module level, but if it's not necessary right now, it might not be a good idea to do it right now (when it will be necessary, the design may have been changed and modifications not valid anymore). 4. the raise should use an exception class of some kind I've made it raise a TypeError... Yes, string based exceptions are very a bad thing ! Attributes are cached while using the managed_attributes() *and* noninherited_managed_attributes() funcs. Those are very similar. I hesitated refactor them, but centralize cache management is probably a good thing. This is the purpose of the second patch, which include this refactoring. As expected, performances decrease. Here's the traditionnal benchmark results: SA 0.3.1, rev 2127: total time 4.29376721382 real0m5.420s user0m4.088s sys 0m0.108s SA with attr. cache, no refactoring (~2X faster, back closed to 0.2.8 speed): total time 2.34819602966 real0m3.013s user0m2.344s sys 0m0.088s SA with attr. cache, with refactoring (30% slower than without refactoring) total time 3.05679416656 real0m3.747s user0m2.988s sys 0m0.068s It's up to you choosing the patch ! IMO, I'm *not* in favor to use refactoring in this case :). Finally, I've put a clear_attribute_cache func which, ... clear the attribute cache. While client code may not have to worry about caching, it may need to clear it... thanks much, this is the kind of user involvement i like. Well, you're very welcome ! Glad to help ! Cheers, Seb -- Sébastien LELONG sebastien.lelong[at]sirloon.net P.S: im beginning to suspect that yield introduces overhead vs. a straight list ? (python interpreter storing stack frames ? dunno). dunnotoo :) --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~--- Index: attributes.py === --- attributes.py (révision 2132) +++ attributes.py (copie de travail) @@ -609,8 +609,18 @@ class AttributeManager(object): allows the instrumentation of object attributes. AttributeManager is stateless, but can be -overridden by subclasses to redefine some of its factory operations. +overridden by subclasses to redefine some of its factory operations. Also be aware AttributeManager +will cache attributes for a given class, allowing not to determine those for each objects (used +in managed_attributes() and noninherited_managed_attributes()). This cache is cleared for a given class +while calling register_attribute(), and can be cleared using clear_attribute_cache() +def __init__(self): +# will cache attributes, indexed by class objects +self._attribute_cache = weakref.WeakKeyDictionary() + +def clear_attribute_cache(self): +self._attribute_cache.clear() + def rollback(self, *obj): retrieves the committed history for each object in the given list, and rolls back the attributes each instance to their original value. @@ -639,21 +649,31 @@ def managed_attributes(self, class_): returns an iterator of all InstrumentedAttribute objects associated with the given class. +if self._attribute_cache.has_key(class_): +return self._attribute_cache[class_] + +self._attribute_cache[class_] = [] if not isinstance(class_, type): -raise repr(class_) + is not a type +raise TypeError(repr(class_) + is not a type) for key in dir(class_): value = getattr(class_, key, None) if
[sqlalchemy] Re: Using pickle to save reflected metadata
I still get the error: Traceback (most recent call last): File U:/ADA_tools/pickle_tables.py, line 27, in ? pickle.dump(metadata_to_pickle,pfile) File C:\Python24\lib\pickle.py, line 1382, in dump Pickler(file, protocol, bin).dump(obj) File C:\Python24\lib\pickle.py, line 231, in dump self.save(obj) File C:\Python24\lib\pickle.py, line 338, in save self.save_reduce(obj=obj, *rv) File C:\Python24\lib\pickle.py, line 433, in save_reduce save(state) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 338, in save self.save_reduce(obj=obj, *rv) File C:\Python24\lib\pickle.py, line 433, in save_reduce save(state) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 338, in save self.save_reduce(obj=obj, *rv) File C:\Python24\lib\pickle.py, line 433, in save_reduce save(state) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 338, in save self.save_reduce(obj=obj, *rv) File C:\Python24\lib\pickle.py, line 433, in save_reduce save(state) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 676, in _batch_setitems save(k) File C:\Python24\lib\pickle.py, line 338, in save self.save_reduce(obj=obj, *rv) File C:\Python24\lib\pickle.py, line 433, in save_reduce save(state) File C:\Python24\lib\pickle.py, line 293, in save f(self, obj) # Call unbound method with explicit self File C:\Python24\lib\pickle.py, line 663, in save_dict self._batch_setitems(obj.iteritems()) File C:\Python24\lib\pickle.py, line 677, in _batch_setitems save(v) File C:\Python24\lib\pickle.py, line 313, in save rv = reduce(self.proto) File C:\Python24\lib\copy_reg.py, line 69, in _reduce_ex raise TypeError, can't pickle %s objects % base.__name__ TypeError: can't pickle module objects --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: guessing sql joins from object level
SA can form joins between tables automatically if the tables express the proper foreign key relationship between each other, and if there is no ambiguity in that relationship; i.e. table A and table B have only one ForeignKeyConstraint (or single ForeignKey) between each other. if you have table A and table B, a join is just: A.join(B) with regards to integrating those joins with mapper queries, see http://www.sqlalchemy.org/docs/datamapping.myt#datamapping_selectrelations_queryjoins . the main keyword argument to select() here is the from_obj parameter. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: ForeignKey and onupdate/ondelete
it was modeled after SQL CREATE TABLE syntax itself, where the ForeignKey keyword looks like column REFERENCES sometable(othercolumn) and does not allow any ON UPDATE or ON DELETE clauses at that level; they would be declared at the bottom of the table def inside the CONSTRAINT foo FOREIGN KEY... line. I suppose I could put onupdate/ondelete into ForeignKey but I am trying to minimize the number of ways to do the same thing. ill try to remember to add this in the next release. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Using pickle to save reflected metadata
latest trunk, this test works for me:
from sqlalchemy import *
meta = BoundMetaData('mysql://user:[EMAIL PROTECTED]/dbname')
table = Table('foo', meta, autoload=True)
meta2 = MetaData()
t2 = table.tometadata(meta2)
import pickle
pickle.dumps(meta2)
post your test case (with some sample table DDLs) and we'll try it out
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: problem with mapper relationship
the test object you are loading in the update function is local to the update function itself. when the function completes, test falls out of scope and is removed from the session (since it is weakly referenced in the session). a change needs to be made to session whereby dirty objects dont get de-referenced from the session (clean objects have to, thats a long-accepted behavior). since this is the second email in three days about this issue, ticket #388 is added. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] count()
Hello, I would like to query a big table and display the information. But at the top of the list I'm going to display, I need a little counter that tell how many items where returned. What would be the most efficient way to count the results ? client_list = session.query(Client).select() Should I: 1/ Use the python len() function: len(client_list) 2/ Use the count() query method. But doesn't it re-query the database ? Thanks for your advices. Regards, -- Alexandre CONRAD --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] assign_mapper breaks polymorphic multi-table inheritance in 3.1?
I'm trying to use assign_mapper with polymorphic multiple table
inheritance, and running into problems with the primary key sequencing.
Is this a supported use of sqlalchemy? I'm running version 0.3.1.
Sample code below. If use_assign_mapper is false, then the script
works with both postgres sqlite. If it's true, then sqlite fails on
the assertion that the manager employee id's be different, and
postgres throws an exception that 'managers_person_id_seq does not
exist'
Thanks in advance for any help, and thanks for an otherwise awesome
product!
-Ken
#!/usr/bin/env python
# The setup here is identical to the example at:
#
http://sqlalchemy.org/docs/adv_datamapping.myt#advdatamapping_inheritance_multiple
# Except that it conditionally uses assignmapper instead of the default
mapper
from sqlalchemy import *
from sqlalchemy.ext.assignmapper import assign_mapper
from sqlalchemy.ext.sessioncontext import SessionContext
use_postgres = False
use_assign_mapper = True
if not use_postgres:
url = 'sqlite:///:memory:'
else:
url=postgres://ken:[EMAIL PROTECTED]/tutorial
engine = create_engine(url)
metadata = BoundMetaData(engine)
context = SessionContext(create_session)
employees = Table('employees', metadata,
Column('person_id', Integer, primary_key=True),
Column('name', String(50)),
Column('type', String(30)))
engineers = Table('engineers', metadata,
Column('person_id', Integer, ForeignKey('employees.person_id'),
primary_key=True),
Column('engineer_info', String(50)),
)
managers = Table('managers', metadata,
Column('person_id', Integer, ForeignKey('employees.person_id'),
primary_key=True),
Column('manager_data', String(50)),
)
class Employee(object):
pass
class Engineer(Employee):
pass
class Manager(Employee):
pass
person_join = polymorphic_union(
{ 'engineer':employees.join(engineers),
'manager':employees.join(managers),
'person':employees.select(employees.c.type=='person'),
}, None, 'pjoin')
metadata.create_all()
###
if not use_assign_mapper:
#This works:
person_mapper = mapper(Employee, employees,
select_table=person_join,
polymorphic_on=person_join.c.type,
polymorphic_identity='person')
mapper(Engineer, engineers, inherits=person_mapper,
polymorphic_identity='engineer')
mapper(Manager, managers, inherits=person_mapper,
polymorphic_identity='manager')
else:
#This doesn't
person_mapper = assign_mapper(context, Employee, employees,
select_table=person_join,
polymorphic_on=person_join.c.type,
polymorphic_identity='person')
assign_mapper(context, Engineer, engineers, inherits=person_mapper,
polymorphic_identity='engineer')
assign_mapper(context, Manager, managers, inherits=person_mapper,
polymorphic_identity='manager')
context.current.echo_uow=True
eng = Engineer()
mgr = Manager()
context.current.save(eng)
context.current.save(mgr)
context.current.flush()
print eng: %d mgr: %d % (eng.person_id, mgr.person_id)
assert(eng.person_id != mgr.person_id)
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: problem with mapper relationship
Michael Bayer ha scritto: the test object you are loading in the update function is local to the update function itself. when the function completes, test falls out of scope and is removed from the session (since it is weakly referenced in the session). a change needs to be made to session whereby dirty objects dont get de-referenced from the session (clean objects have to, thats a long-accepted behavior). since this is the second email in three days about this issue, ticket #388 is added. Thanks. This is a crucial feature for me, since I use this pattern in Twisted (where in a separate thread I create a transaction and a session, passing the connection and session objects to an user define function). Regards Manlio Perillo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: count()
if you are going to fetch all results in all cases, then len(result) is fine. if the result list is enormous and youre only going to fetch a small portion of it within a request, the count() query as a separate operation is better. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Using pickle to save reflected metadata
whoops...
I've been using IDLE to test this and I guess it has to be reopened in
order for imports to be redone (I was just closing the shell window).
pickle.dump(metadata) works fine now.
But now I get a (recursive?) error when I try to unpickle (and I have
to kill the shell):
import pickle
pfile = file('pickled_tables','rb')
metadata = pickle.load(pfile)
Traceback (most recent call last):
File pyshell#7, line 1, in -toplevel-
metadata = pickle.load(pfile)
File C:\Python24\lib\pickle.py, line 1390, in load
return Unpickler(file).load()
File C:\Python24\lib\pickle.py, line 872, in load
dispatch[key](self)
File C:\Python24\lib\pickle.py, line 1235, in load_build
setstate = getattr(inst, __setstate__, None)
File C:\Python24\lib\site-packages\sqlalchemy\util.py, line 113, in
__getattr__
return self.__data[key]
File C:\Python24\lib\site-packages\sqlalchemy\util.py, line 113, in
__getattr__
return self.__data[key]
File C:\Python24\lib\site-packages\sqlalchemy\util.py, line 113, in
__getattr__
.
.
.
keeps on going
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: PickleType with custom pickler
On Dec 8, 2006, at 1:30 PM, Michael Bayer wrote: the reason that get_history calls attribute_manager.get_history is because all awareness of attribute history is handled by the attribute module. the reason attribute_manager then calls the InstrumentedAttribute off of the class is because thats where AttributeManager reprents the behavior for a specific class attribute. ColumnProperty knows nothing about InstrumentedAttribute. Hmm...I guess ColumnProperty is not really what I'm looking for then-- it looks more like a layer between user code and the attribute manager. I want to be able to control the communication between the attribute manager and the mapped object. Currently AttributeManager and friends just access object.__dict__ directly, which means I can't get between them...too bad. I guess what I'm really looking to customize is InstrumentedAttribute, which is even more formidable than ColumnProperty. Is there an easy way to tell SA to use a custom InstrumentedAttribute class? if you look at the constructor for InstrumentedAttribute itself, youll see it takes two callables which provide the compare function and the copy function. these are provided by the TypeEngine object via the ColumnLoader, which is off the ColumnProperty. So i would propose that these two arguments become optional keyword arguments to ColumnProperty which ColumnLoader/DeferredLoader pick up on and send to their init_class_attribute method, so that you can send in whatever compare and copy callables you want. I'm not sure if that would actually solve my problem since compare and copy do not allow me to mutate the data as it passes from SA to the object. Could we create a mapper extension point that would get called to check if an object is modified? It would need be hooked into the logic inside locate_dirty(). That would allow me to do this whole thing with a mapper extension. ~ Daniel --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: PickleType with custom pickler
while youre free to make your own InstrumentedAttribute, i think that is a long uphill battle which also might be incompatible with new releases. I think you might want to go with sticking some hidden attributes on your pickleable object so that your custom PickleType can just use those to identify the information it needs (a good choice might be the id of the session, so that you can get the session back by saying session._sessions[id]). or, perhaps the item gets a reference back to its parent so it can find the session via object_session(parent). or you could decouple it by using a weak-keyed dictionary somewhere so that your pickle instances can be mapped back to parent and therefore session. this might be something you can set up within a MapperExtension using populate_instance() when instances are loaded and before_insert() when instances are about to be first saved. parent_dict = weakref.WeakValueDictionary() class MyExt(MapperExtension): def populate_instance(self, mapper, selectcontext, row, instance, identitykey, isnew): if not isnew: return EXT_PASS mapper.populate_instance(context, instance, row, identitykey, isnew) parent_dict[instance.my_pickled_thing] = instance return None class MyPickleType(MapperExtension): def convert_bind_param(self, value, dialect): return MyPickler(session = object_session(parent_dict[value])).dumps(value) ... --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: SQL Alchemy
On 12/4/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Can you use real numbers and date into SQL Alchemy database? In a word: yes. For real numbers, you'd use the Float() column type, and for dates, you'd use the Date() column type. (Or the DateTime() column type if you wanted both a date and time value). http://www.sqlalchemy.org/docs/types.myt has a list of the basic data types available, and an example of how to write your own types should one of those not be sufficient. http://www.sqlalchemy.org/docs/metadata.myt shows how to declare a table. Hopefully this will answer your question. -- Robin Munn [EMAIL PROTECTED] GPG key 0x4543D577 --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: remote nondirect access to DB
Jonathan Ellis píše v Pá 08. 12. 2006 v 11:05 -0700: On 12/7/06, ml [EMAIL PROTECTED] wrote: I want to have a client application accessing a remote Postgres database but I don't want to distribute the user/password to the DB. I want clients to authenticate against some other table of passwords so there must be a stub between client and DB which will know the DB user/pwd. How can I manage this in SA? You can write your stub in SA easily; your client should interface with the stub not via SQL but via a simple protocol you define that limits it to predetermined actions. How? What protocol? Any clues or examples are welcome. DS --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: remote nondirect access to DB
On 12/8/06, ml [EMAIL PROTECTED] wrote: Jonathan Ellis píše v Pá 08. 12. 2006 v 11:05 -0700: On 12/7/06, ml [EMAIL PROTECTED] wrote: I want to have a client application accessing a remote Postgres database but I don't want to distribute the user/password to the DB. I want clients to authenticate against some other table of passwords so there must be a stub between client and DB which will know the DB user/pwd. How can I manage this in SA? You can write your stub in SA easily; your client should interface with the stub not via SQL but via a simple protocol you define that limits it to predetermined actions. How? What protocol? Any clues or examples are welcome. You get to make one up! It's your app! The point is, if you trust the client to send valid SQL you are screwed whether or not the client logs into the database directly or not. So instead of insert into users (name) values (...) your client would sent NEWUSER ... or something. If this sounds intimidating, I would suggest picking up a book covering client/server programming. I believe Programming Python and Core Python both include chapters on socket programming. -- Jonathan Ellis http://spyced.blogspot.com --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: PickleType with custom pickler
this is a simple example of what youre trying to do. its not even as
fancy as what i described above and just uses a threadlocal to keep
track of the current Session:
from sqlalchemy import *
from sqlalchemy.orm.session import object_session
from cStringIO import StringIO
from pickle import Pickler, Unpickler
import threading
meta = BoundMetaData('sqlite://', echo=True)
class MyExt(MapperExtension):
def populate_instance(self, mapper, selectcontext, row, instance,
identitykey, isnew):
# if you access the row before this step, it fails
MyPickler.sessions.current = selectcontext.session
return EXT_PASS
def before_insert(self, mapper, connection, instance):
MyPickler.sessions.current = object_session(instance)
return EXT_PASS
def before_update(self, mapper, connection, instance):
MyPickler.sessions.current = object_session(instance)
return EXT_PASS
class MyPickler(object):
sessions = threading.local()
def persistent_id(self, obj):
if hasattr(obj, id):
key = %s:%s % (type(obj).__name__, obj.id)
return key
return None
def persistent_load(self, key):
print KEY, key
name, ident = key.split(:)
sess = MyPickler.sessions.current
return sess.query(Bar).get(ident)
def dumps(self, graph, protocol):
src = StringIO()
pickler = Pickler(src)
pickler.persistent_id = self.persistent_id
pickler.dump(graph)
return src.getvalue()
def loads(self, data):
dst = StringIO(data)
unpickler = Unpickler(dst)
unpickler.persistent_load = self.persistent_load
return unpickler.load()
foo_table = Table('foo', meta,
Column('id', Integer, primary_key=True),
Column('bar', PickleType(pickler=MyPickler()), nullable=False))
bar_table = Table('bar', meta,
Column('id', Integer, primary_key=True),
Column('data', String(40)))
meta.create_all()
class Foo(object):
pass
class Bar(object):
def __init__(self, value):
self.data = value
mapper(Foo, foo_table, extension=MyExt())
mapper(Bar, bar_table)
sess = create_session()
b = Bar('some bar')
sess.save(b)
sess.flush()
sess.clear()
sess = create_session()
f = Foo()
f.bar = b
sess.save(f)
sess.flush()
sess.clear()
# delete current session to make sure its populated
del MyPickler.sessions.current
f = sess.query(Foo).get(f.id)
assert f.bar.data == 'some bar'
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
sqlalchemy group.
To post to this group, send email to sqlalchemy@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~--~~~~--~~--~--~---
[sqlalchemy] Re: remote nondirect access to DB
Jonathan Ellis píše v Pá 08. 12. 2006 v 15:40 -0700: On 12/8/06, ml [EMAIL PROTECTED] wrote: Jonathan Ellis píše v Pá 08. 12. 2006 v 11:05 -0700: On 12/7/06, ml [EMAIL PROTECTED] wrote: I want to have a client application accessing a remote Postgres database but I don't want to distribute the user/password to the DB. I want clients to authenticate against some other table of passwords so there must be a stub between client and DB which will know the DB user/pwd. How can I manage this in SA? You can write your stub in SA easily; your client should interface with the stub not via SQL but via a simple protocol you define that limits it to predetermined actions. How? What protocol? Any clues or examples are welcome. You get to make one up! It's your app! The point is, if you trust the client to send valid SQL you are screwed whether or not the client logs into the database directly or not. So instead of insert into users (name) values (...) your client would sent NEWUSER ... or something. If this sounds intimidating, I would suggest picking up a book covering client/server programming. I believe Programming Python and Core Python both include chapters on socket programming. Now I see where are you heading. So the idea is to put all SA into the stub and each action will be exposed via a single method. It is not what I wanted but it seems I have no other choice. Thanks. DS --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Proposal: session identity_map no longer weak referencing
On 12/8/06, Michael Bayer [EMAIL PROTECTED] wrote: when SA was first released, someone immediately suggested that the identity map of Session be weak referencing, which appeared to be an obvious improvement, so that you could load as many objects as you want from the session and whatever you didnt use would just go away. but now it appears that the more intuitive operation for a Session is that things that get loaded into it, stay there, until you say otherwise. So id like to make the session act this way, again like hibernate's...more explicit. what you load into the session stays there until you clean it out. that behavior is more predictable than what we have now. +1 Of course we can always add an option weak_identity_map if people really do want the old behavior. -1, it's only a matter of time before this would bite you -- Jonathan Ellis http://spyced.blogspot.com --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sqlalchemy group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
