On Jan 4, 11:56 am, mdipierro <[email protected]> wrote:
> > I'm using google's raw datastore api (google.ext.db). I gave up on
> > web2py's style of models when I really needed a ListProperty of
> > KeyProperty's.
>
> Have you checked out the latest IS_IN_DB(....multiple=True)?
> I do not think it is very different than the ListProperty.
In all fairness, I have not. I have seen posts discussing it, and
I've taken a quick glance at the code.
It's more a gut reaction based on [very limited] experience and
exposure to issues I see on the GAE mailing list. Well, those and a
lack of time.
IS_IN_DB seems to solve other issues than the ones I've run across,
though (I could very well be misunderstanding its use).
In one example, I have an ACL of users/groups allowed to do various
things. Under each security condition, I keep a ListProperty of User/
Group keys. If one of those keys matches the currently logged-in user
(or any of the groups the user's in), the user can do whatever that
permission level allows. I'm leaning toward storing the user's ACL
rights as a pickled BLOB instead (in a record corresponding to each
user), but, for now, this is performing acceptably.
I switched to doing it that way a few months back, because I couldn't
figure out how to do it at all with web2py's ORM.
In another example, I have a tree data structure. Child nodes contain
reference nodes to their parent nodes. The parent node has a
BooleanProperty to indicate whether it has any children (this is used
in the tree view to indicate whether a given node can be expanded or
not). Since ReferenceProperty's work both ways, I can get the Node's
children in a single query when the client tries to expand a node.
Before I switched to that method, App Engine timed out all over the
place.
I'm tempted to de-normalize that as well, but I also query directly
for child nodes, and I want to avoid the duplication as much as I can.
Anyway, looking through the newest sources, it looks like the id
column has been switched to use GAE's key, so maybe these use-cases
could be converted back without much overhead.
Please don't anyone read this as any sort of criticism of web2py's
ORM. Far from it. The fact that web2py runs on GAE out of the box
puts it way ahead of any comparable framework (including Google's old
crippled version of django) that I've been able to find.
I almost always wind up putting a wrapper layer over any data access
layer I use, almost out of habit. I've had to change those so many
times that I don't even really think twice about that particular
step. I made the engineering trade-off decision to switch to google's
ORM a few months back, because it made the most sense for that
situation at that time for that current project.
Based upon Sudhakar.M's description, I suspected that he'd run into
many of the same issues, so I recommended he run some tests and
benchmarks to see.
(One of the GAE trade-offs, as far as I've been able to tell, is that
performance doesn't seem to vary much with the size of your data
store, though it does with the size of your results. So querying a
few thousand records should give you a fairly accurate benchmark. I
really should check that assumption).
I have found that working directly with their API has given me a much
better feel for what works and what doesn't. It's easy enough to say
that it isn't an RDBMS. It's a lot harder to break years' worth of
habits of thinking in RDBMS terms. (I still see at least 2-3 posts a
week on their mailing list of new developers complaining that queries
time out, but they aren't willing to de-normalize their data).
Actually, thinking about it from that perspective, it seems like
another good reason to not use web2py's ORM. If/lwhen I do switch to
a different platform, I'll also be switching back to an RDBMS. Which
will involve a fairly hefty database re-design, no matter which way I
look at things. I'll be far less likely to miss something if I have
an obvious re-write where anything I miss will break immediately.
Sorry for the long babble. These are issues that I keep circling back
and questioning myself about.
Regards,
James
>
> Massimo
>
> On Jan 4, 11:37 am, James Ashley <[email protected]> wrote:
>
> > [This is starting to veer *way* off-topic, and might be more
> > appropriate on the GAE list. My apologies]
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"web2py Web Framework" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/web2py?hl=en
-~----------~----~----~----~------~----~------~--~---
