a urgent question from a new comer
Dear All I have a newbie but urgent question about OJB. Each db in my project might finally contain around 1 million objects defined by more than 500 classes. The classes structure is terrifically complex with many levels of inheritance and references. I might query the db using some ODMG API like, a.b.c.d.e.f.g, which is seven levels of references. The final results might contains a few thousands of objects. I can wait for a few second for a query but not a few minutes. I am not sure that OJB is suitable for my project. Does anybody have any experience with large quantity of data with a complex class structure? Best regards Cui Lu __ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
GUI Mapping Tool
Hi, I am developping a gui mapping tool for OJB since some time. I search some people interested to develop this tool with me. Developpers who know well OJB, the advanced OO concepts, design patterns, UML, Swing or have the motivation to develop this cool software are welcome. Let me know if you are interested. Jean Mori
Re: [Re-post] PB query question
hi phil,
there have been discussions about defining the fields of the subclasses
in the abstract base class as well. but i do not know the current state :(
you could try to use a path hint (setPathForClass) but this could
conflict with the other extent .
jakob
Phil Warrick wrote:
Hi again,
I posted this a few weeks ago, perhaps when the best person to answer
wasn't around (Jakob?). Your suggestions would be welcome.
Here is my model:
[A] <1--n> [B] <1--n> [C] <1--1> [D]
| |
C1 C2
I want to query for A objects using a C1 timestamp field as criteria (C1
is a subclass of C and C does not have the timestamp field) and a D
string field as criteria.
If the timestamp field belonged to the C class, then there would be no
problem directly querying for A's.
Since the criteria is a C1 attribute I don't think that I can directly
query for A's, so I can do it in two steps by first querying for C1
objects and then iterating through the returned collection to gather the
A objects, as follows
// snippit begin
Criteria criteria1 = new Criteria();
Criteria criteria2 = new Criteria();
criteria1.addEqualTo("d.stringField", new String("foo"));
criteria2.addEqualTo("d.stringField", new String("foofoo"));
criteria1.addOrCriteria(criteria2);
Criteria criteria3 = new Criteria();
criteria3.addBetween("timeStamp", timeStamp1, timeStamp2);
criteria1.addAndCriteria(criteria3);
Query query = new QueryByCriteria(C1.class, criteria1);
Collection c1Set = broker.getCollectionByQuery(query);
java.util.Iterator itr = c1Set.iterator();
List aList = new ArrayList();
while (itr.hasNext())
{
// gather the A's (more db hits here)
A a = c1.getB().getA();
aList.add(a);
}
// snippit end
It would be nice to reduce the DB hits and query for the A's directly as
is possible in with standard OQL:
select c1.b.a from c1 in C1.class
where (c1.d.stringField = \"foo\" or
c1.d.stringField = \"foofoo\") and
e.theTimestamp >= timestamp '" + timeStamp1 + "' " and
e.theTimestamp <= timestamp '" + timeStamp2 + "' "
Is there any possibility that this is possible with PB queries?
Thanks for read this far,
Phil
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: Need JOIN
Hi Alen,
OJB will handle joins for you. Try something like the following.
crt.addEqualTo("id", new Integer(3));
crt.addEqualTo("allProductsForCategory.activate", new Boolean(true));
Query query = new QueryByCriteria(query, crt, true); //DISTINCT
Collection results = broker.getCollectionByQuery(queryByCriteria);
-Wally
-Original Message-
From: Alen Ribic [mailto:[EMAIL PROTECTED]
Sent: Thursday, August 21, 2003 4:38 PM
To: OJB Users List
Subject: Need JOIN
Hi All
I have a single Category object.
This Category object contains a collection of Products of type Vector.
Product objects contain a boolean field named 'activate'.
I have a method that returns a single category with all products in it's
category by category PK.
Everything is fine as is but as soon as I try to specify a product
specific
criteria, I get an exception that such column does not exist. Column I
wish
to use as my criteria (WHERE clause) is 'activate'.
I was thinking that the category would be joining onto all applicable
products but that doesn't seem to be the case. It must be executing
separate
sql statements.
I which to use criteria:
query = ".id = 3 AND .activate
= 1"
as follows:
...
crt.addSql(query);
QueryByCriteria queryByCriteria =
QueryFactory.newQuery(Category.class, crt, true);
Collection results = broker.getCollectionByQuery(queryByCriteria);
my repository 's:
Any help, thx
--Alen
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Need JOIN
Hi All I have a single Category object. This Category object contains a collection of Products of type Vector. Product objects contain a boolean field named 'activate'. I have a method that returns a single category with all products in it's category by category PK. Everything is fine as is but as soon as I try to specify a product specific criteria, I get an exception that such column does not exist. Column I wish to use as my criteria (WHERE clause) is 'activate'. I was thinking that the category would be joining onto all applicable products but that doesn't seem to be the case. It must be executing separate sql statements. I which to use criteria: query = ".id = 3 AND .activate = 1" as follows: ... crt.addSql(query); QueryByCriteria queryByCriteria = QueryFactory.newQuery(Category.class, crt, true); Collection results = broker.getCollectionByQuery(queryByCriteria); my repository 's: Any help, thx --Alen - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Question re prefetched relationsbips
Here's the patch and a new file. When putting in the new code I tried to just
splice it in gently so it would be easy to update from CVS later and also
would not affect OJB at all if you don't use the functionality, so it might
not be integrated quite as nicely as you'd expect. As I mentioned the first
time I submitted this I don't think it handles extents correctly, and maybe
some other things, but it does work quite well for what I'm using it for.
This app has one object per table and mostly just plain fk references. See
the messages linked to below for more info. If I have time I'd be happy to
help work on any of these issues.
To use, you'd use a QueryByCriteriaMultiJoin instead of regular query. The
"extraJoin"s you add are relationship path names, any depth. Querying on A,
with a relationship to B named "relB", where B has one to C ("relC"), you
could add extra joins for "relB" and "relB.relC". The query would return all
fields from tables a, b, and c and build the objects B and C that A
references. When accessing a.getB().getC() the objects are already loaded and
there's no extra queries - just the one original query on A. This works best
with one-to-one relationships, or if you select from the many side of a
one-to-many. You just have to be aware of what's going on since there's going
to be table joining going on in order to get the extra fields.
Hope that makes sense. I just updated 5pm CST 2003-08-21 and that's what the
diff is based on.
John Marshall
Connectria
>= Original Message From Jakob Braeuchi <[EMAIL PROTECTED]> =
>hi john,
>
>could you please send it to me as well ?
>
>jakob
>
>John M wrote:
>
>>I submitted a patch earlier that allows a different method of prefetching
>>related objects and does indeed allow prefetching for more than one level
>>deep. All data is returned in one query by joining tables and getting all
>>fields for the related objects. The patch files were based off what I got
out
>>of cvs and aren't based on the tagged rc4, so I'll have to update the patch
>>for you if you are interested.
>>
>>I think deep prefetching is useful for our app, which at times needs to load
a
>>large set of data for some analysis and optimization and the whole object
>>graph is useful. It is also useful for many reports.
>>
>>The patch was included on this message:
>>http://nagoya.apache.org/eyebrowse/[EMAIL PROTECTED]&m
sgId=752859
>>
>>and described here:
>>http://nagoya.apache.org/eyebrowse/[EMAIL PROTECTED]&m
sgId=751991
>>
>>
>>John Marshall
>>Connectria
>>
>>
>>
>>>= Original Message From "OJB Users List" <[EMAIL PROTECTED]> =
>>>Hello,
>>>
>>>
>>>
>>>-Original Message-
>>>From: Jakob Braeuchi [mailto:[EMAIL PROTECTED]
>>>
>>>
>>>prefetching currently works only for one level . the prefetcher looks
>>>for a relationship named
>>>
>>>"allb.singlec" in A.class in your sample. the query criteria
>>>has no effect on prefetching.
>>>i'm not sure whether deep prefetching is really useful.
>>>
>>>
>>>Such a feature would really make me happy. What do you think,
>>>is it difficult to implement? Are there other people
>>>having that need?
>>>
>>>Olli
>>>
>>>
>>>
Does anyone know if prefetched relationships can reach more
>>>than one "layer"
>>>
>>>
of references/collections from an class being queried?
E.g. If I have a class A that I an querying and it has a 1:m
>>>relationship
>>>
>>>
with some class B, with the relationship defined by the name "allb",
and then each B has 1 1:1 relationsbip with another class C, with the
relationship called "singlec", is the following supposed to work:
Criteria criteria = new Criteria();
// configure the criteria
...
// add prefetched relationships
criteria.addPrefetchedRelationship("allb");
criteria.addPrefetchedRelationship("allb.singlec");
Query q = QueryFactory.newQuery( A.class, criteria);
join-load.zip
Description: Zip compressed data
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Serialize result set / collection
Hi, I'm using OJB and I want to save a result set to file, then recreate it at a later time. I want to do something like the following: Collection result = broker.getCollectionByQuery(query); File file = new File(filename); OutputStream out = new FileOutputStream(file); MySerializer ser = new MySerializer(); ser.serialize(out, result); out.close(); (here the user can pause from minutes to several days/weeks) Then at a later point in time I want to recreate the collection from the file. File file = new File(filename); InputStream in = new FileInputStream(file); MySerializer ser = new MySerializer(); Collection result = ser.unserialize(in); in.close(); Iterator itor = result.iterator(); while(itor.hasNext()) System.out.println(itor.next()); BUT, I do not want to serialize the whole object(s). I just want to write the primary key of each object in the result set to the file. Before I start to re-invent the wheel (which I have done many times), is there something similar already? If not, how do I get hold of each primary key in a collection in an efficient way (without materializing each object)? Somewhere in the back of my head I heard of a java standard for "disconnected result sets". Is that something I should dig deeper into? /Goran Ehrsson - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
ODMG behavior with multiple databases and multiple clients
Hello everybody,
We are using the ODMG layer with OJB RC4 to connect to 2 DB2 databases and 1
Oracle database. We found out that each time we open a database through ODMG
the previous one is closed. This behavior is coded in the OJB class. It
seems that only the current default database can be open at any time.
Additionally the transaction interface does not allow to define the database
on which it will act on.
Examples from commercial ODMG products show that it should be possible to
specify the database on which a transaction should work on and have multiple
databases open. Currently we redefined the OJB class to support the creation
of a new transaction with a Parameter indicating the database on which to
execute.
Configurator configurator =
PersistenceBrokerFactory.getConfigurator();
...
// This method is an extension of the method defined in
org.apache.ojb.odmg.OJB
public Transaction newTransaction(Database database) {
TransactionImpl tx = new
TransactionImpl((DatabaseImpl)database);
try {
configurator.configure(tx);
} catch (ConfiguratorException e) {
...
}
The multiple databases are managed as follow.
// workaround of the problem that Database.open calls
OJB.registerOpenDatabase(DatabaseImpl) which always close the previously
opened database.
public synchronized Database getOracleDatabase() throws
ODMGException {
if (oracleDatabase == null) {
oracleDatabase = odmg.newDatabase();
oracleDatabase.open(ORACLE_DATABASE_REPOSITORY,
Database.OPEN_READ_WRITE);
} else if (!(DatabaseImpl)oracleDatabase).isOpen()) {
oracleDatabase.open(ORACLE_DATABASE_REPOSITORY,
Database.OPEN_READ_WRITE);
}
return oracleDatabase;
}
The above solution cannot work in an environment where multiple clients
access the server and get various databases in parallel for obvious reasons.
What is the correct way to access simultaneously multiple database through
the ODMG layer using OJB RC4?
Thanks for any pointer to a FAQ or an answer. We did not found an example in
the OJB FAQ or the Mail archives. Examples for commercial ODMG products show
that multiple databases can be opened At the same time and that a
transaction can be created with a database parameter.
What are we doing wrong or have missed?
Marcel Baumann
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Sub query
hi, please have a look at the posts 'repost: subqueries' jakob Dougall Squair wrote: Hi, Is it possible to have a sub-querey within OJB that referers to its enclosing query? - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem inserting object when using SequenceManagerNativeImpl
Hi Armin,
I've done some more debugging and it seems the problem is that
SequenceManagerNativeImpl.afterStore() always gets called, even if the
class does not have an autoincrement field. The reason I did not have
this problem previously is that I was using a custom version of
SequenceManagerMySQLImpl instead (sorry I had forgotten about this). I
changed SequenceManagerNativeImpl.afterStore() to the following and all
is working now:
public void afterStore(JdbcAccess dbAccess, ClassDescriptor cld,
Object obj) throws SequenceManagerException
{
FieldDescriptor fd = cld.getAutoIncrementField();
if (fd == null) {
if (log.isDebugEnabled()) {
log.debug("Class does not have an autoincrement field,
skipping: " + cld);
}
} else {
long newId = getLastInsert(cld);
setFieldValue(obj, fd, new Long(newId));
}
}
Thanks,
Mike
On Wednesday, August 20, 2003, at 07:23 PM, Armin Waibel wrote:
> Hi Michael,
>
>> - Inserting an object that does not have an autoincrement field
> hmm, changes were made to SequenceManagerNativeImpl last 10 days,
> but no changes in detecting autoincrement fields. Each pk field must
> be set before storing or need an autoincrement attribute (set 'true')
> in
> field-descriptor (default was 'false').
>
> regards,
> Armin
>
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: repost: subqueries
Hello Jakob, > -Original Message- > From: Jakob Braeuchi [mailto:[EMAIL PROTECTED] > please see my post in dev-list '[vote] reference to a parent > attribute > in subquery' > this allows you to reference attributes of the enclosing > query. please > be aware that this is a _temporary_ solution, the prefix may change ! looks great! Thank you, I will try that! BTW, I recommend to make that prefix a public String constant, i.e. PARENT_QUERY_PREFIX and encourage to use that constant, then its value is of minor importance. Olli - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: JDO inheritance problem
On Thu, 21 Aug 2003 08:54:07 -0500 Ketan Gangatirkar <[EMAIL PROTECTED]> wrote: > persistence-capable-superclass="A" That's what I was missing ! Thank you ! I went a little further and ... got another (smaller, I think) problem : java.lang.ClassCastException at org.apache.ojb.jdori.sql.OjbExtent.(Unknown Source) at org.apache.ojb.jdori.sql.OjbStoreManager.getExtent(Unknown Source) at com.sun.jdori.common.PersistenceManagerImpl.getExtent(Unknown Source) at com.sun.jdori.common.query.QueryImpl.checkCandidates(Unknown Source) at com.sun.jdori.common.query.QueryImpl.execute(Unknown Source) at UCListAllStatutes.apply(UCListAllStatutes.java:38) ... Statue2 get recorded in its own table correctly. Statute also But I can't list Statute the way I used to : manager.currentTransaction().begin(); Query query = manager.newQuery(Statute.class); Collection allStatutes = (Collection)query.execute(); The last line give me the exception. Any idea about that ? Cédric -- Societe Code Lutin 2, rue Robert le Ricolais BP 20446 44304 NANTES CEDEX 3 Tel 02.40.50.29.28, Fax 02.40.93.99.21 http://www.codelutin.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: OJB Console
Yes, I am sorry. I am using SourceForge to host my project, but the "demo" is being hosted from my computer out of my homeso be gentle :) I have been having a few issues with it and hotspot errors. I think I need to recompile my kernel. I'm sorry for that. I started everything back up for the time being, but I will work on correcting it. -alexander You wrote: > Alexander Bibighaus wrote: > >You can demo the application here: > >http://www.scrashmeow.org/ojb/index.do > > > Internal Server Error > The server encountered an internal error or misconfiguration and was > unable to complete your request. > Please contact the server administrator, [EMAIL PROTECTED] and inform > them of the time the error occurred, and anything you might have done > that may have caused the error. > More information about this error may be available in the server error log. > > Apache/2.0.44 (Unix) PHP/4.3.1 mod_jk/1.2.1 Server at www.scrashmeow.org > Port 80 > - > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] -- -alexander - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: JDO inheritance problem
Cédric,
You need to define the attribute persistent-capable-superclass in your
.jdo file for the sub-class. So if you have class A extends class B,
your .jdo file would have:
The JDORI enhancer creates final methods to access the persistent
attributes of the object. If it does not know that B is a subclass of
A, it will define them for both B and A, which will result in the error
you see. If A has an attribute z, then you should list it only in the
section of A. Since B is a subclass of A, it will inherit the
getter/setters that JDORI creates for z. Only define B's unique
attributes in the section for B. Like so:
Cédric Pineau wrote:
Hello,
I'm trying to use OJB through JDO and have a problem with basic
inheritance between two persitent class :
I get "
class entities.Statute2 overrides final method .
java.lang.VerifyError: class entities.Statute2 overrides final
method .
at java.lang.ClassLoader.defineClass0(Native Method)
... "
at runtime when the classloader loads the Statute2 class.
Statute2 inherits from Statute (this actually a test drive :-) Here is
my code :
Statute :
package entities;
public class Statute {
protected int id;
public int getId() {
return this.id;
}
public void setId(int id) {
this.id = id;
}
/** default constructor */
public Statute() {}
/** full constructor */
public Statute(java.lang.String name) {
this.name = name;
}
protected java.lang.String name;
public java.lang.String getName() {
return this.name;
}
public void setName(java.lang.String name) {
this.name = name;
}
public String toString() {
return "Statute "+this.name+" (id = "+this.id+")";
}
}
Statute 2 :
package entities;
public class Statute2 extends Statute {
/** default constructor */
public Statute2() {}
/** full constructor */
public Statute2(java.lang.String name, String name2) {
this.name = name;
this.name2 = name2;
}
protected java.lang.String name2;
public java.lang.String getName2() {
return this.name2;
}
public void setName2(java.lang.String name2) {
this.name2 = name2;
}
public String toString() {
return "Statute "+this.name+" "+this.name2+" (id = "+this.id+")";
}
}
repository_user.xml :
primarykey="true" autoincrement="true"/>
primarykey="true" autoincrement="true"/>
entities.jdo (partial) :
...
...
sql for table creation (postgresql) :
CREATE TABLE STATUTE (
ID INT PRIMARY KEY,
NAME TEXT
);
CREATE TABLE STATUTE2 (
ID INT PRIMARY KEY,
NAME TEXT,
NAME2 TEXT
);
I suppose I did something wrong but can't figure what.
I already tried to enhance only one of the two, to remove id and name
from Statute2 class-desciptor entry but...
Can someone light my way to the wonderful world of JDO ?
Thanks,
Cédric
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
JDO inheritance problem
Hello,
I'm trying to use OJB through JDO and have a problem with basic
inheritance between two persitent class :
I get "
class entities.Statute2 overrides final method .
java.lang.VerifyError: class entities.Statute2 overrides final
method .
at java.lang.ClassLoader.defineClass0(Native Method)
... "
at runtime when the classloader loads the Statute2 class.
Statute2 inherits from Statute (this actually a test drive :-) Here is
my code :
>> Statute :
package entities;
public class Statute {
protected int id;
public int getId() {
return this.id;
}
public void setId(int id) {
this.id = id;
}
/** default constructor */
public Statute() {}
/** full constructor */
public Statute(java.lang.String name) {
this.name = name;
}
protected java.lang.String name;
public java.lang.String getName() {
return this.name;
}
public void setName(java.lang.String name) {
this.name = name;
}
public String toString() {
return "Statute "+this.name+" (id = "+this.id+")";
}
}
>> Statute 2 :
package entities;
public class Statute2 extends Statute {
/** default constructor */
public Statute2() {}
/** full constructor */
public Statute2(java.lang.String name, String name2) {
this.name = name;
this.name2 = name2;
}
protected java.lang.String name2;
public java.lang.String getName2() {
return this.name2;
}
public void setName2(java.lang.String name2) {
this.name2 = name2;
}
public String toString() {
return "Statute "+this.name+" "+this.name2+" (id = "+this.id+")";
}
}
repository_user.xml :
entities.jdo (partial) :
...
...
sql for table creation (postgresql) :
CREATE TABLE STATUTE (
ID INT PRIMARY KEY,
NAME TEXT
);
CREATE TABLE STATUTE2 (
ID INT PRIMARY KEY,
NAME TEXT,
NAME2 TEXT
);
I suppose I did something wrong but can't figure what.
I already tried to enhance only one of the two, to remove id and name
from Statute2 class-desciptor entry but...
Can someone light my way to the wonderful world of JDO ?
Thanks,
Cédric
--
Societe Code Lutin
2, rue Robert le Ricolais
BP 20446 44304 NANTES CEDEX 3
Tel 02.40.50.29.28, Fax 02.40.93.99.21
http://www.codelutin.com
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: OJB Console
Alexander Bibighaus wrote: You can demo the application here: http://www.scrashmeow.org/ojb/index.do Internal Server Error The server encountered an internal error or misconfiguration and was unable to complete your request. Please contact the server administrator, [EMAIL PROTECTED] and inform them of the time the error occurred, and anything you might have done that may have caused the error. More information about this error may be available in the server error log. Apache/2.0.44 (Unix) PHP/4.3.1 mod_jk/1.2.1 Server at www.scrashmeow.org Port 80 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
