Thread: [Beankeeper-discussion] BeanKeeper instance identity.
Brought to you by:
demonsystem
Menu
▾
▴
beankeeper-discussion
|
From: Sergey O. <ser...@we...> - 2011-12-08 15:31:53
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body text="" bgcolor="">
Hi,<br>
<br>
I am not sure if this list is even alive anymore, but let's see :)<br>
<br>
Recently I've been trying to use BeanKeeper. It seemed to cover
mostly what I need and is very simple to use -- which is great!<br>
<br>
However now I discovered something that has me completely confused.
I assumed that BeanKeeper manages Java object instances -- to give a
really trivial example -- if I query for something by its unique ID
twice, I'd expect to receive the same Java object instance both
times. However this not appears to be the case?<br>
<br>
For example, if I execute the following twice in a row:<br>
log.info("Offer lookup result HASH: " +
System.identityHashCode(DbService.findSingle("find DbOffer where
DbOffer.offerId = '" + offer.getId() + "'")));<br>
I get different hash codes.<br>
<br>
This seems very weird to me -- how do I work with a rich data model
when I can't count on the same object to be represented by the same
instance? For example, I have two different paths to the same
instance via Lists and Maps -- and it seems that I'll be working
with different instance depending on which path I take -- and if I
use both paths, it'll probably result in a mess.<br>
<br>
Am I missing something here?<br>
<br>
Best regards,<br>
Sergey<br>
</body>
</html>
|
|
From: Marcel S. <ma...@fr...> - 2011-12-09 19:17:42
|
On 08.12.11 16:31, Sergey Olefir wrote:
> Hi,
>
> I am not sure if this list is even alive anymore, but let's see :)
I didn't get a message in years. So yeah, I'll be interested myself.
> Recently I've been trying to use BeanKeeper. It seemed to cover mostly
> what I need and is very simple to use -- which is great!
>
> However now I discovered something that has me completely confused. I
> assumed that BeanKeeper manages Java object instances -- to give a
> really trivial example -- if I query for something by its unique ID
> twice, I'd expect to receive the same Java object instance both times.
> However this not appears to be the case?
I don't know for sure but I wouldn't be surprised if what you're seeing
is the expected behavior. I realize that there may be use cases where
such a behavior would be helpful but AFAIK you can't count on it with
any ORM tool.
With Hibernate for example you could achieve what you want only if use a
second level cache (if I'm not mistaken).
> For example, if I execute the following twice in a row:
> log.info("Offer lookup result HASH: " +
> System.identityHashCode(DbService.findSingle("find DbOffer where
> DbOffer.offerId = '" + offer.getId() + "'")));
> I get different hash codes.
>
> This seems very weird to me -- how do I work with a rich data model when
> I can't count on the same object to be represented by the same instance?
> For example, I have two different paths to the same instance via Lists
> and Maps -- and it seems that I'll be working with different instance
> depending on which path I take -- and if I use both paths, it'll
> probably result in a mess.
>
> Am I missing something here?
>
> Best regards,
> Sergey
>
>
> ------------------------------------------------------------------------------
> Cloud Services Checklist: Pricing and Packaging Optimization
> This white paper is intended to serve as a reference, checklist and point of
> discussion for anyone considering optimizing the pricing and packaging model
> of a cloud services business. Read Now!
> http://www.accelacomm.com/jaw/sfnl/114/51491232/
>
>
>
> _______________________________________________
> Beankeeper-discussion mailing list
> Bea...@li...
> https://lists.sourceforge.net/lists/listinfo/beankeeper-discussion
--
Marcel Stör, http://www.frightanic.com
Couchsurfing: http://www.couchsurfing.com/people/marcelstoer
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org
|
|
From: Sergey O. <so...@gm...> - 2011-12-09 20:03:48
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body text="" bgcolor="">
<br>
With Hibernate everything you load within the scope of one Session
(which for e.g. web application you'd usually set to span the entire
request processing) is guaranteed to have one Java instance per one
database object / row -- i.e. no matter how you load the object,
it'll resolve to the same Java instance.<br>
<br>
<a class="moz-txt-link-freetext" href="http://docs.jboss.org/hibernate/core/3.3/reference/en/html/transactions.html">http://docs.jboss.org/hibernate/core/3.3/reference/en/html/transactions.html</a><br>
"For objects attached to a particular Session (i.e., in the scope of
a Session), the two notions are equivalent and JVM identity for
database identity is guaranteed by Hibernate."<br>
<br>
The same is true for general JPA. This is basically has been a
cornerstone on which Hibernate / JPA is built. That's why I was so
surprised it's not apparently true for BeanKeeper.<br>
<br>
For what it worth, I've done a bit of code digging and I'm going to
bet that indeed in BeanKeeper if you reach a database object
instance / row via different paths (via different query or by
following a different set of collections) -- you'll have a different
Java instance unlike traditional ORM.<br>
<br>
It also seems that it would be trivially easy to add some kind of
Session support to BeanKeeper, but it cannot be done externally
(requires modification of BeanKeeper's code directly).<br>
<br>
However after actually trying to develop minor application on
BeanKeeper I found other limitations and shortcomings --
particularly inability to have attribute with the same name in two
classes if they share a common parent and incompatibility with most
recent HSQLDB. So it seems a serious risk to invest into BeanKeeper
as it shows that it hasn't been developed/maintained for a while.
But given how much I like it conceptually, I might still give it a
shot.<br>
<br>
It's a pity it hadn't gained a following and seems to be 'mostly
dead' for now as it seems it truly has serious potential as a proper
way to do objects persistence in Java without all the mess
associated with even good ORM (Hibernate) to say nothing of a worse
solutions (EJB pre version 3 anyone?).<br>
<br>
On 09.12.2011 21:02, Marcel Stör wrote:
<blockquote cite="mid:4EE...@fr..." type="cite">
<pre wrap="">On 08.12.11 16:31, Sergey Olefir wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Hi,
I am not sure if this list is even alive anymore, but let's see :)
</pre>
</blockquote>
<pre wrap="">
I didn't get a message in years. So yeah, I'll be interested myself.
</pre>
<blockquote type="cite">
<pre wrap="">Recently I've been trying to use BeanKeeper. It seemed to cover mostly
what I need and is very simple to use -- which is great!
However now I discovered something that has me completely confused. I
assumed that BeanKeeper manages Java object instances -- to give a
really trivial example -- if I query for something by its unique ID
twice, I'd expect to receive the same Java object instance both times.
However this not appears to be the case?
</pre>
</blockquote>
<pre wrap="">
I don't know for sure but I wouldn't be surprised if what you're seeing
is the expected behavior. I realize that there may be use cases where
such a behavior would be helpful but AFAIK you can't count on it with
any ORM tool.
With Hibernate for example you could achieve what you want only if use a
second level cache (if I'm not mistaken).
</pre>
<blockquote type="cite">
<pre wrap="">For example, if I execute the following twice in a row:
log.info("Offer lookup result HASH: " +
System.identityHashCode(DbService.findSingle("find DbOffer where
DbOffer.offerId = '" + offer.getId() + "'")));
I get different hash codes.
This seems very weird to me -- how do I work with a rich data model when
I can't count on the same object to be represented by the same instance?
For example, I have two different paths to the same instance via Lists
and Maps -- and it seems that I'll be working with different instance
depending on which path I take -- and if I use both paths, it'll
probably result in a mess.
Am I missing something here?
Best regards,
Sergey
------------------------------------------------------------------------------
Cloud Services Checklist: Pricing and Packaging Optimization
This white paper is intended to serve as a reference, checklist and point of
discussion for anyone considering optimizing the pricing and packaging model
of a cloud services business. Read Now!
<a class="moz-txt-link-freetext" href="http://www.accelacomm.com/jaw/sfnl/114/51491232/">http://www.accelacomm.com/jaw/sfnl/114/51491232/</a>
_______________________________________________
Beankeeper-discussion mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Bea...@li...">Bea...@li...</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/beankeeper-discussion">https://lists.sourceforge.net/lists/listinfo/beankeeper-discussion</a>
</pre>
</blockquote>
<pre wrap="">
</pre>
</blockquote>
</body>
</html>
|
|
From: Marcel S. <ma...@fr...> - 2011-12-09 20:08:12
|
On 09.12.11 21:03, Sergey Olefir wrote: > With Hibernate everything you load within the scope of one Session > (which for e.g. web application you'd usually set to span the entire > request processing) is guaranteed to have one Java instance per one > database object / row -- i.e. no matter how you load the object, it'll > resolve to the same Java instance. Right, as long as the session-scope is what you're after that's correct. I thought your requirement was to get the same object for consecutive requests, sorry. Cheers, Marcel -- Marcel Stör, http://www.frightanic.com Couchsurfing: http://www.couchsurfing.com/people/marcelstoer O< ascii ribbon campaign - stop html mail - www.asciiribbon.org |
|
From: Bräutigam R. <de...@ne...> - 2011-12-13 19:46:14
|
On 12/08/2011 04:31 PM, Sergey Olefir wrote: > Hi, > > I am not sure if this list is even alive anymore, but let's see Hi, The list is alive and monitored, however BeanKeeper is not actively developed anymore. I still think its approach is a valid one, but unfortunately all the developers (the few we've had) moved on to other projects, including myself. > Recently I've been trying to use BeanKeeper. It seemed to cover mostly > what I need and is very simple to use -- which is great! Glad to hear that. > However now I discovered something that has me completely confused. I > assumed that BeanKeeper manages Java object instances -- to give a > really trivial example -- if I query for something by its unique ID > twice, I'd expect to receive the same Java object instance both times. > However this not appears to be the case? You are correct, this is not the case, and on purpose. The reason is that BeanKeeper tries to be as simple as possible even if that means it does not fit all use-cases. BeanKeeper doesn't have any notion of a "Session" (of course it does have a notion of a Transaction though). It doesn't manage bean instances for you like Hibernate does. I believe for somebody not coming from Hibernate (which should be really rare these days), this is actually more intuitive. The rule is: any object you receive is completely clear, and does not in any way have any side-effects. It does not persist automatically, it does not influence other instances in your JVM, etc. > Best regards, > Sergey Robert. |
|
From: Sergey O. <so...@gm...> - 2011-12-14 09:53:11
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="" bgcolor="">
First of all -- thanks for replying!<br>
<br>
I can understand the logic behind not wanting to do Session
mechanism -- even if I don't necessarily agree with it (because it
makes working with persistent objects completely different from
working with POJOs).<br>
<br>
With that said, I have a question. Let's take a look at the
following pseudo-code (which is something I actually encountered
trying to use BeanKeeper):<br>
tx<br>
{<br>
a = find... // retrieve persistent object from BeanKeeper, it
has 'list' property with list containing 2 elements.<br>
instance1 = a.list.get(1);<br>
instance1.update()<br>
instance1.save()<br>
instance2 = a.list.get(0); // Before 1 above -- this is
important!<br>
instance2.update()<br>
instance2.save()<br>
instance3 = a.list.get(1); // instance3 != instance1 at this
point!!!<br>
instance3.update2()<br>
instance3.save()<br>
}<br>
txcommit<br>
<br>
As the result of all this I find that changes made in
instance1.update() (to the instance1 fields) are NOT persisted. I'm
guessing here that reading list backwards causes BeanKeeper to
discard and recreate get(1) element because of performance
optimizations (where it reads several items at once rather than just
one). This in turn means that code operates on two different Java
instances for get(1) and attempts to change and save both -- with
changes made to instance1 subsequently lost.<br>
<br>
What I'm wondering about -- is this considered a bug or a feature?
If it is a feature, what is the rule-of-thumb for working with
BeanKeeper that'll prevent me from losing updates like the example
above?<br>
<br>
Best regards,<br>
Sergey<br>
<br>
On 13.12.2011 21:46, Bräutigam Róbert wrote:
<blockquote cite="mid:4EE...@ne..." type="cite">
<pre wrap="">On 12/08/2011 04:31 PM, Sergey Olefir wrote:
> Hi,
>
> I am not sure if this list is even alive anymore, but let's see
Hi,
The list is alive and monitored, however BeanKeeper is not actively
developed anymore.
I still think its approach is a valid one, but unfortunately all the
developers (the few we've had) moved on to other projects, including myself.
> Recently I've been trying to use BeanKeeper. It seemed to cover mostly
> what I need and is very simple to use -- which is great!
Glad to hear that.
> However now I discovered something that has me completely confused. I
> assumed that BeanKeeper manages Java object instances -- to give a
> really trivial example -- if I query for something by its unique ID
> twice, I'd expect to receive the same Java object instance both times.
> However this not appears to be the case?
You are correct, this is not the case, and on purpose. The reason is
that BeanKeeper tries to be as simple as possible even if that means it
does not fit all use-cases.
BeanKeeper doesn't have any notion of a "Session" (of course it does
have a notion of a Transaction though). It doesn't manage bean instances
for you like Hibernate does. I believe for somebody not coming from
Hibernate (which should be really rare these days), this is actually
more intuitive.
The rule is: any object you receive is completely clear, and does not in
any way have any side-effects. It does not persist automatically, it
does not influence other instances in your JVM, etc.
> Best regards,
> Sergey
Robert.
------------------------------------------------------------------------------
Systems Optimization Self Assessment
Improve efficiency and utilization of IT resources. Drive out cost and
improve service delivery. Take 5 minutes to use this Systems Optimization
Self Assessment. <a class="moz-txt-link-freetext" href="http://www.accelacomm.com/jaw/sdnl/114/51450054/">http://www.accelacomm.com/jaw/sdnl/114/51450054/</a>
_______________________________________________
Beankeeper-discussion mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Bea...@li...">Bea...@li...</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/beankeeper-discussion">https://lists.sourceforge.net/lists/listinfo/beankeeper-discussion</a>
</pre>
</blockquote>
</body>
</html>
|
Thanks for helping keep SourceForge clean.
X