[sqlalchemy] ORM __eq__

2012-02-15 Thread Michael Hipp 

Doing this to test equality between ORM objects:

  a = sess1.query(Thing).get(1)
  b = sess2.query(Thing).get(1)  # different session

Was somewhat surprised to find that:
  a == bgives False

Is this by design? Do I need to add a custom __eq__ method to my declarative 
Base classes to make a==b come out True?


Thanks,
Michael Hipp

--
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 
sqlalchemy+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sqlalchemy?hl=en.

Re: [sqlalchemy] ORM __eq__

2012-02-15 Thread Michael Bayer 

On Feb 15, 2012, at 12:54 PM, Michael Hipp wrote:

 Doing this to test equality between ORM objects:
 
  a = sess1.query(Thing).get(1)
  b = sess2.query(Thing).get(1)  # different session
 
 Was somewhat surprised to find that:
  a == bgives False
 
 Is this by design? Do I need to add a custom __eq__ method to my declarative 
 Base classes to make a==b come out True?

Curious, for a and b to be equals without a custom __eq__, that would 
imply that sess1 and sess2 share the same identity map and state, is that what 
you were expecting ?   Or just that SQLAlchemy would automatically instrument 
an __eq__() method on mapped classes ?

Anyway, yeah we don't assume that, so you'd want to implement an __eq__() that 
I'd assume compares primary key values.


-- 
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 
sqlalchemy+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sqlalchemy?hl=en.

Re: [sqlalchemy] Usage of begin nested

2012-02-15 Thread Michael Bayer 

On Feb 15, 2012, at 12:21 AM, Manav Goel wrote:

 My use case requires that if insertion of object of Myclass succeeds
 then insert Object of Myclass1 .
 Even if inserting of Myclass1 object fails insertion of Myclass should
 not be rolled back.
 I mean adding Myclass is permanent and does not depend on failure or
 success of insertion of Myclass1.
 I have written following code and want to know if am understanding
 right usage of begin_nested and not writing buggy code.
 I am using postgresql 9.0
 try:
 obj =Myclass()
 session.add(obj)
 if condition true:
session.begin_nested()
try:
n = Myclass1(arguments)
session.add(n)
except SQLAlchemyError:
db_session.rollback()
 
  session.commit()
 except SQLAlchemyError:
  session.rollback()
  raise
 
 Code is running f9, just want to make sure of some unknown gotcha in
 this code.
 Other option will be I commit after adding Myclass and perform
 insertion of Myclass1 in separate transaction but this way is not
 appealing to me.

assuming that db_session is a typo for session, then yes this is the basic 
idea.There's a slight gotcha in that you're not inspecting the actual 
SQLAlchemyError coming in, you might want to limit that to IntegrityError which 
is what psycopg2 usually throws for these.

-- 
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 
sqlalchemy+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sqlalchemy?hl=en.

Re: [sqlalchemy] Full Table scan with Oracle due to Charset conversion

2012-02-15 Thread Michael Bayer 

On Feb 15, 2012, at 3:39 PM, Harish Tejwani wrote:

 Need help!
 
 SQL Alchemy: 0.6.8
 cx_Oracle: 5.1.1
 Red-had Linux
 Oracle 11gR2, NLS_CHARACTERSET AL32UTF8
 
 Simple Queries like
 
 SELECT events.source_name AS events_source_name
 FROM events
 WHERE events.source_mac_id = :source_mac_id_1
 
 2012-02-15 10:07:13,061 INFO sqlalchemy.engine.base.Engine.0x...9410
 {'source_mac_id_1': u'9911A'}
 
 Bind Variables has a 'u' prefixed that causes Oracle to do full TS as
 it uses SYS_OP_C2C to do conversion and ignores the index on columns.
 
 class Event(Base):
'''
'''
 
__tablename__ = events
 
source_name = Column(BigInteger)
source_mac_id = Column(String(128))
 
 ..
 
 Tried all settings on engine connect side and also changing as below,
 still no help

SQLAlchemy never converts to a u'', however if this is data that's coming back 
from cx_oracle as a u'' being passed back again, that might be the cause.If 
you don't pass in a u'' string, then it will remain as a plain bytestring.   

If you want to force it to encode even when a u'' is passed, use 
convert_unicode='force':

String(128, convert_unicode='force')

Though this really seems like a cx_oracle issue.  It shouldn't matter if u'' or 
'' is passed, the end result should be similar and this seems like some kind of 
Oracle mis-configuration.   You might want to ask on the cx_oracle list.




-- 
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 
sqlalchemy+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sqlalchemy?hl=en.