UpdateAllQuery

I get an exception when I run an updateAllQuery without setting a SelectionCriteria.
Is it not possible to do a whole table update using updateAllQuery?
Can someone please point what is wrong.
Thanks
UpdateAllQuery updateQuery = new UpdateAllQuery(Person.class);
updateQuery.addUpdate("name", "Z");
updateQuery.setShouldDeferExecutionInUOW(false);
updateQuery.setCacheUsage(UpdateAllQuery.INVALIDATE_CACHE);
uow.executeQuery(updateQuery);
Exception:
Exception [TOPLINK-6089] (Oracle TopLink - 10g Release 3 (10.1.3.1.0) (Build RELEASE)): oracle.toplink.exceptions.QueryException
Exception Description: The expression has not been initialized correctly. Only a single ExpressionBuilder should be used for a query.
For parallel expressions, the query class must be provided to the ExpressionBuilder constructor, and the query's ExpressionBuilder must
always be on the left side of the expression.

Thanks for the update.
Is there any other way to add 1=1?
I am using this
updateQuery.setSelectionCriteria(updateQuery.getExpressionBuilder().postfixSQL("1=1"));
It works, but the JavaDoc mentions this -
Warning: Allowing an unverified SQL string to be passed into this method makes your application vulnerable to SQL injection attacks.

Similar Messages

AboutToUpdate and UpdateAllQuery

We have extended DescriptorEventAdaptor to use the aboutToUpdate method for auditing fields. That works fine.
When an UpdateAllQuery is executed on an object I would expect the aboutToUpdate to add the auditing fields in the update statement too, specially when the locking field(version in my test case) gets added to my update statement.
But my tests show that this is not the case
The update statement that’s gets generated by UpdateAllQuery is -
UPDATE USER04.COMMODITY_RELATIONX SET VERSION = (VERSION + ?), DESCRIPTION = UPPER(DESCRIPTION) WHERE (NAME = ?)
Can someone please explain this and let me know what I am doing something wrong here.
Thanks.
This is my test –
ExpressionBuilder builder = new ExpressionBuilder();
Expression WhereClause = builder.get("name").equal("UpdateTest");
UpdateAllQuery updateQuery = new
UpdateAllQuery(CommodityRelationX.class);
updateQuery.setSelectionCriteria(WhereClause);
updateQuery.addUpdate(builder.get("description"),
builder.get("description").getFunction("UPPER"));
updateQuery.setShouldDeferExecutionInUOW(false);
int num = (Integer)uow.executeQuery(updateQuery);
txn.commit();

In your event first determine that this query is the UpdateAllQuery that you want, then add your updates.
If the column are not mapped you could use a query-key, or can use a field expression,
updateQuery.addUpdate(updateQuery.getExpressionBuilder().getField("UPDATEDBYCOL"), "UpdateUser");

Toplink UpdateAllQuery Issue

We're experiencing some problems with an UpdateAllQuery when trying to update some fields belonging to an aggregate mapping within the component we're trying to update.
We want to set these fields to NULL.
One of the fields in the Aggregate mapping is a OneToOne mapping lookup type component. We want to set this field to NULL. (In a nutshell we want to set the FK to this lookup type object to NULL).
In order to accomplish this, we have created a DirectQueryKey for the OneToOne mapping inside of the Aggregate mapping and then physically mapped it within the Component that makes use of the Aggregate mapping.
It appears that this DirectQueryKey is not working properly. Even when we create a ReadAllQuery and use the same selection criteria, TopLink complains with a TOPLINK 6015 "Invalid QueryKey [businessFuncId] in expression" exception when trying to execute the query.
We have checked to make sure the QueryKey name in our TopLink mapping matches the name used in the query definition... we're stumped... can someone review and provide some clues as to what we may be doing wrong.
Here is the UpdateAllQuery definition:
UpdateAllQuery updateQuery = new UpdateAllQuery(RegisterImpl.class);
updateQuery.setCacheUsage(UpdateAllQuery.INVALIDATE_CACHE);
ExpressionBuilder registerBuilder = updateQuery.getExpressionBuilder();
updateQuery.addArgument("userId");
updateQuery.addArgument("channel");
updateQuery.addArgument("token");
updateQuery.addArgument("date");
updateQuery.addArgument("businessFuncId");
Expression reservedDetailReg = registerBuilder.get("reservedDetail");
Expression userIdExp = reservedDetailReg.get("userId").equal(
registerBuilder.getParameter("userId"));
Expression channelExp = reservedDetailReg.get("channel").equal(
registerBuilder.getParameter("channel"));
Expression tokenExp = reservedDetailReg.get("token").equal(
registerBuilder.getParameter("token"));
Expression dateExp = reservedDetailReg.get("date").equal(
registerBuilder.getParameter("date"));
Expression busFuncExp = reservedDetailReg.get("businessFuncId").equal(
registerBuilder.getParameter("businessFuncId"));
updateQuery.setSelectionCriteria(userIdExp.and(channelExp.and(tokenExp.and(dateExp
.and(busFuncExp)))));
updateQuery.addUpdate(reservedDetailReg.get("userId"), "");
updateQuery.addUpdate(reservedDetailReg.get("channel"), "");
updateQuery.addUpdate(reservedDetailReg.get("token"), "");
updateQuery.addUpdate(reservedDetailReg.get("date"), "");
updateQuery.addUpdate(reservedDetailReg.get("businessFuncId"), "");
descriptor.getQueryManager().addQuery(QueryConstants.UNFREEZE_REGISTER_BY_RESERVED_DETAIL,
updateQuery);
Thanks in advance.

In answer to your posted questions:
1) Are you mapping a query key for the lookup object(between an attribute of the object and a field of the table) or are you adding a query key for the relationship mapping?
Answer: We are mapping a DirectQueryKey (see "businessFuncId" above) to a field on the table directly related to the component being updated. However, the query key represents the foreign key (primary key) for a OneToOne mapping contained within an aggregate descriptor.
Here's an explanation of the OR mapping/object relationship (reference the query definition in the original posting for more details):
RegisterImpl (target of updateAllQuery) contains a ReservedDetail (Aggregate mapping) which in turn contains a OneToOne mapping to a BusinessFunction (read-only lookup type object).
We created a DirectQueryKey: businessFuncId which represents the FK on the RegisterImpl table that is used within the ReservedDetail aggregate mapping for the OneToOne mapping to the BusinessFunction component.
This query key is for query purposes only. As you can see from the query definition it is used to avoid a join with a BusinessFunction table. This is because, as we understand it, an UpdateAllQuery can only update ONE target component (table). The problem we're having is: how do you set a OneToOne FK reference field on the target component to NULL???
2) To which descriptor are you adding the query key?
Answer: The query key ("businessFuncId") must be, by convention, defined within the ReservedDetail aggregate mapping. When we map the ReservedDetail aggregate mapping within the RegisterImpl component mapping, we map the query key to the businessFunction FK field contained with the RegisterImpl table, (the target table we're trying to update).

UpdateAllQuery and Native SQL

Hi,
I have below three update statements, and need to be in one transaction. How can I accomplish it using UpdateAllQuery? Finance_Management is a child table of Org mapped through org_id; Org_Finance is a child table of Finance_Management linked by finance_management_id.
Update Org
Set Account_Balance = Account_Balance + 25, Update_User = 'Jeff_Org', Update_Date = Systimestamp
Where Org_Id = 10226;
Update Finance_Management
Set Account_Balance = (Select Account_Balance From Org Where Org_Id = 10226),
Transaction_Amount = 25, Transaction_Date = Systimestamp
where finance_management_id = 1701152;
Update Org_Finance
Set Check_Mo_Number = 5432, Payment_Type = 'C'
where Finance_Management_Id = 1701152;
Alternatively, when I tried native SQL with executeNonSelectingCall within unit of work like below, it appears that even through the second SQLCall failed, the first one still committed to the database. How can I ensure the three SQLCalls are in one transaction?
UnitOfWork uow = getSessionFactory().acquireUnitOfWork();
uow.executeNonSelectingCall(new SQLCall(orgSql));
uow.executeNonSelectingCall(new SQLCall(fmSql));
uow.executeNonSelectingCall(new SQLCall(orgFnSql));
uow.commit();
Thanks for your help!
Jeffrey

I think I have figured it out.
For using UpdateAllQuery, three different UpdateAllQueries need to be constructed. One critical piece is that, use beginEarlyTransaction() before execute the three queries inside UOW. Same is true when using native SQLs.
Few questions though that I need some experts to clear my mind:
- so all database Queries can run inside unit of work, become part of the single transaction within UOW by using beginEarlyTransaction(), is this correct?
- How about insert? InsertObjectQuery and WriteObjectQuery don't seem allow to use expression, which I need for populating a field like "Account_Balance = (Select Account_Balance From Org Where Org_Id = 10226)". Any suggestions on what API call I should use without using native SQL?
Thanks,
Jeffrey

Toplink Cache not getting refreshed after executing UpdateAllQuery

After executing UpdateAllQuery, the records in the database are getting updated, however the Toplink cache is not getting refreshed with the new data, it still has stale data which is causing issues.
We're also setting
updateQuery.setCacheUsage(UpdateAllQuery.INVALIDATE_CACHE);
Thanks.

Toplink version is 10.1.3.0.0
Here is the code
UpdateAllQuery updateQuery = new UpdateAllQuery(RegisterImpl.class);
updateQuery.setCacheUsage(UpdateAllQuery.INVALIDATE_CACHE);
ExpressionBuilder registerBuilder = updateQuery.getExpressionBuilder();
updateQuery.addArgument("userIdArg");
updateQuery.addArgument("channelArg");
updateQuery.addArgument("tokenArg");
updateQuery.addArgument("dateArg");
updateQuery.addArgument("businessFuncIdArg");
Expression reservedDetailRegExp = registerBuilder.get("reservedDetail");
// build expressions
Expression userIdExp = reservedDetailRegExp.get("userId").equal(
registerBuilder.getParameter("userIdArg"));
Expression channelExp = reservedDetailRegExp.get("channel").equal(
registerBuilder.getParameter("channelArg"));
Expression tokenExp = reservedDetailRegExp.get("token").equal(
registerBuilder.getParameter("tokenArg"));
Expression dateExp = reservedDetailRegExp.get("date").equal(
registerBuilder.getParameter("dateArg"));
Expression busFuncExp = registerBuilder.get("businessFuncId").equal(
registerBuilder.getParameter("businessFuncIdArg"));
// set selection criteria
updateQuery.setSelectionCriteria(userIdExp.and(channelExp.and(tokenExp.and(dateExp
.and(busFuncExp)))));
// substitute the values
updateQuery.addUpdate(reservedDetailRegExp.get("userId"), "");
updateQuery.addUpdate(reservedDetailRegExp.get("channel"), "");
updateQuery.addUpdate(reservedDetailRegExp.get("token"), "");
updateQuery.addUpdate(reservedDetailRegExp.get("date"), "");
updateQuery.addUpdate(registerBuilder.get("businessFuncId"), "");
In the object model for the query, the RegisterImpl has an aggregate mapping (ReservedDetail), which in turn has a number of direct-to-field mappings and an one-to-one mapping to businessFunction (for our query we use a direct query key "businessFuncId")

Bulk Insert/Update in Toplink 10.1.3.

Hi experts,
I want to update a column in the database for all rows of a pariticular table. Also want the change to reflect in Toplink cache on all nodes of the Weblogic cluster. The cache in various nodes are synchronized using a JMS Topic. I want to avoid registering all these objects in the UnitOfWork
for performance reasons. The changes do not seem to propogate when I use other Bulk update methods. Is there a standard way of doing this?
Thanks,
Kamal

You can update a set of rows using an Update All JPQL query in JPA, or using the native UpdateAllQuery class.
An Update All query will invalidate the local cache, but is not currently broadcast across cache coordination.
The Cache/IdentityMapAccessor invalidateObject() API allows an object to be invalidate across the cluster, but not a class or query set.
Please log a bug for this on EclipseLink, and contact Oracle technical support if you need a patch for this.
James : http://www.eclipselink.org

Aggregate descriptors and query keys

We are in the process of migrating from TopLink to EclipseLink.
I have a class Audit that contains the auditing columns that are common to other classes
I have another class Customer that extends the Audit.
Audit is an aggregate descriptor mapped in Customer.
There is an UpdateAllQuery to update a single customer row in the customer.
To update the audit column an addUpdate is added on the field.
query.addUpdate(builder.getField("LAST_UPDT_USER_C"), "Test");
When this query is executed I get an error saying the field is not mapped. But this field is mapped in the aggregate.
This used to work in TopLink.
Any idea why it errors out and how to solve it?
Thanks for the help.
The JavaDoc says that getField work on non mapped field -
Expression org.eclipse.persistence.internal.expressions.DataExpression.getField(String fieldName)
ADVANCED: Return an expression representing a field in a data-level query. This is used internally in EclipseLink, or to construct queries involving fields and/or tables that are not mapped.
ERROR -
Exception Description: Attribute name or expression passed as a first parameter to addUpdate method defines a field from a table that's not mapped to query descriptor.
Attribute name or Expression: [
Field LAST_UPDT_USER_C
Base QUERY OBJECT]
Wrong field: [LAST_UPDT_USER_C]

Hello,
I can't see what has changed in the area that throws the exception, as it seems similar to TopLink 11. Which version of TopLink was this working on, and what version of EclipseLink are you using?
If the field is mapped, why are you not using the attribute mapping via the builder.get("audit").get("auditMappingForTheField") instead of accessing the field directly?
The code itself seems to throw an exception because the table for the field cannot be found in the descriptor's table list You can try to define the table name in the field and see if this helps: builder.getField("Tablename.LAST_UPDT_USER_C"). If this works, please file a bug as it should not be neccessary or the error message needs to be fixed.
Best Regards,
Chris

Toplink compared to its competitors and futur of this product in O/R domain

I am sincerely disappointed by the evolution of TopLink whereas I have been in favour of this product for several years at important customers.
1- New version 10.1.3 (10.0.3 included) has been in beta for almost 3 years!
2- patch corrects only the bugs without any evolution on the level of the functionalities, all the new functionalities are in the 10.1.3 which does not manage to leave. By way of comparison of the product as KODO JDO brings in each patch: New features + Notable Exchanges + Bugfixes 3.4.0 - 3.3.4 - 3.3.3 - 3.3.2 - 3.3.1 - 3.3.0 -! I AM JEALOUS OF KOKO CUSTOMERS
You can see KOKO evolution on 2005 at : http://www.solarmetric.com/Software/Documentation/3.4.0/docs/relnotes.html
3- Other editors (KODO+COCOBASE+JBOSS) start to set up beta versions EJB 3.0 with very very complete documenatation and product. Whereas Toplink provides only examples (tutorial) over OC4J (10.1.3 preview in beta since 3 years!) . No real toplink (bata vversion) of Documentation or Product. Like KODO 4 @ : http://www.solarmetric.com/Software/beta/4.0.0EA/
+ Real and Rich documentation @ : http://www.solarmetric.com/Software/Documentation/4.0.0EA/docs/
4- 4 10.1.3 preview Toplink brings functionalities JCA, JAXP but not much of things on the level of mapping O/R: updateAll, cacheEviction, returning, java.util.logging. These functionalities already available on the others concurrent product do not make me any more dream !
I hope that it other evolutions (not announced) on Object/Relational scope ?
5- does Oracle have still wants to really invest in O/R mapping. There is much marketing, much of mediatic show (Eclipse, GlassFish, EJB3 lead) but not much of technical evolutions compared to competitors.
Thank you to product management make a factual answer because I am a real customer, I am really in favour of the product for 5 years. But I am really disappointed when I see new small actors (KODO...) much better doing?

I am sincerely disappointed by the evolution of
TopLink whereas I have been in favour of this product
for several years at important customers. I am sorry to hear about your disappointment. Hopefully I can provide some insight into where we are with TopLink. It is a critical part of our J2EE/EJB infrastructure providing its classic POJO persistence as well as EJB 2.1 CMP/BMP, and now our preview support of EJB 3.0 persistence.
1- New version 10.1.3 (10.0.3 included) has been in
beta for almost 3 years! As an integrated part of the Oracle Application Server we ship in coordination with the overall release. I understand your frustration and we are working hard to ensure we have a high quality 10.1.3 AS release as well as a first rate persistence solution.
2- patch corrects only the bugs without any evolution
on the level of the functionalities, all the new
functionalities are in the 10.1.3 which does not
manage to leave. By way of comparison of the product
as KODO JDO brings in each patch: New features +
Notable Exchanges + Bugfixes 3.4.0 - 3.3.4 - 3.3.3 -
3.3.2 - 3.3.1 - 3.3.0 -! I AM JEALOUS OF KOKO
CUSTOMERS
You can see KOKO evolution on 2005 at :
http://www.solarmetric.com/Software/Documentation/3.4.
0/docs/relnotes.htmlOur patch-sets are just that. Patches to fix issues important to our customers. They are not new feature releases. As an independent product (prior to the Oracle acquisition) we also had many smaller releases but we were also not well integrated with any container. This integration providing better diagnostics, management, and out of the box support requires us to integrate with the container. I believe you will find Kodo will start to feel this as BEA starts to deliver an integrated persistence solution that may compete with TopLink's complete solution.
Note: This does not mean that we are not supporting other containers. We do extensive testing across non-Oracle databases and application servers to ensure that we continue to work well there as well.
3- Other editors (KODO+COCOBASE+JBOSS) start to set
up beta versions EJB 3.0 with very very complete
documenatation and product. Whereas Toplink provides
only examples (tutorial) over OC4J (10.1.3 preview in
beta since 3 years!) . No real toplink (bata
vversion) of Documentation or Product. Like KODO 4 @
: http://www.solarmetric.com/Software/beta/4.0.0EA/
+ Real and Rich documentation @ :
http://www.solarmetric.com/Software/Documentation/4.0.
0EA/docs/At the time of our EJB 3.0 preview we had the best support available for both within the container as well as outside the container. This was primarily documented through how-to's. As we move towards our release (post 10.1.3) that implements the complete specification, which is not yet final, we will be delivering more complete and formal documentation.
The feedback I had received on our how-to's has been excellent to date. It does require more updates now with our 10.1.3 release to address additional functionality we will be providing in the EJB 3.0 persistence areas.
4- 4 10.1.3 preview Toplink brings functionalities
JCA, JAXP but not much of things on the level of
mapping O/R: updateAll, cacheEviction, returning,
java.util.logging. These functionalities already
available on the others concurrent product do not
make me any more dream !
I hope that it other evolutions (not announced) on
Object/Relational scope ?If you look at the list of new features in 10.1.3 here you will see that we have also invested heavily in improvements to our core ORM.
updateAll = UpdateAllQuery
cacheEviction = cache invalidation/expiration (time-to-live) as well as cache coordination mode using invalidation
returning= we have a returning policy for use with triggers and stored procs
logging = we have updated our logging infrastructure to use java.util.logging
Many of these and others have been available in our earlier previews as well. We have a dedicated team here focused both on ORM as well as OXM (including JCA). I would estimate that 75% of our efforts have been on enhanced ORM functionality.
>
5- does Oracle have still wants to really invest in
O/R mapping. There is much marketing, much of
mediatic show (Eclipse, GlassFish, EJB3 lead) but not
much of technical evolutions compared to
competitors.I believe with 10.1.3 we raise the bar of ORM persistence solutions. The new ORM functionality and improved mapping editors will provide existing users new features they have been asking for as well as help grow the developer community. Our community edition of TopLink in GlassFish delivers core ORM persistence to those projects where the full licensed product is not required as well as providing these customers a seamless upgrade to the full solution when they wish to.
As this release comes out we will be updating our collateral and I will strive to ensure our continued leadership and investment in ORM functionality is reflected.
>
Thank you to product management make a factual answer
because I am a real customer, I am really in favour
of the product for 5 years. But I am really
disappointed when I see new small actors (KODO...)
much better doing?I hope you find this helpful. I would be more than happy to discuss this further. This does make it clear that we need to be more vocal and clear about our work as well as getting the 10.1.3 release out the door. I appreciate the feedback and will try to address these concerns directly.
Doug Clarke
Principal Product Manager
Oracle TopLink

Update Query using Toplink

I would like to fire an update query using Toplink mapping objects. The
structure of the Query would be like
update <<table name>>
set <<col name A>> = 'Y'
where <<col name B>> in (<<valA>>,<<valB>>............);
ValA, ValB.... are dynamic values passed from my front end Java App
Can anyone help me in this regard...........
Thanks
Karthik

Karthik,
This is new functionality added in 10.1.3. You will want to look at UpdateAllQuery.
Doug

Inserting into table whith select query...

Hi All,
I am having requirement where
INSERT into employee_Temp (emp_id,emp_name,emp_email) (select emp_id,emp_name,emp_email from EMP where sal > 300) order by emp_name.
Here EMP table has Toplink mapping to a java object.So I have constructed the select part of the query using ReportQuery.
Can you help me in constructing the complete query so that employee_Temp will be populated in one shot.
I am using Toplink Toplink9.0.3
Thanks in Advance,
Ashwin B

TopLink does not currently support an InsertAllQuery, (10.1.3 added an UpdateAllQuery, and DeleteAllQuery will be supported in 11.1.1).
You will need to use SQL for your query.

Mass update

Hi All,
have two important questions:
In TopLink, is there an efficient way to do an UPDATE statement for say 10,000 rows, and still not use a stored procedure? Something like "update all instruments for this organization to rating AAA," but without the overhead of looping through a collection of instruments.
In TopLink, how do we insert a row and initialize it to a current database timestamp? Not Java timestamp, but database. Something like "INSERT INTO table (modified) VALUES (CURRENT TIMESTAMP)"?
Please help!

In the 10g (9.0.4) version mass updates are only supported through custom SQL. We are adding support for these through the TopLink query and expression framework for 10g (10.0.3). This UpdateAllQuery will optimize these sorts of operations while still ensuring cached instances are properly updated or invalidated.
Again, the upcoming 10.0.3 release will support the returning clause that will allow these values to be set in triggers and returned to Toplink and the application in the result of the INSERT. Currently this can only be accomplished using a stored procedure.
Doug

Update Statement with variable

Hi,
We have an Update statement which sets the account balance as shown in below example:
update Acct_Balance_Table set Balance = Balance + 500 where Acct_Id = 1234
where "Balance" is the balance column name.
Is there a way we can accomplish this using TopLink API, or we have to do it in native SQL?
Thanks for your help!
Jeffrey

One more question.
If I do a insert like below, what TopLink API should I use? UpdateAllQuery seems only for update right? I was looking at WriteObjectQuery, but could not figure out how to add expression for the select sub-query. Any suggestion?
insert into finance_management (transaction_date, transaction_amount, account_balance)
VALUES (SYSTIMESTAMP, 500,
(SELECT account_balance+500 from org_balance where org_id=1234))
Also, if I put both UpdateAllQuery and WriteObjectQuery inside the unit of work, will uow handle both as one transaction?
Thanks,
Jeffrey

Using Toplink to generate UPDATE sql statements?

I have a need to generate a UPDATE sql statement.
This statement will run at startup and will not effect the cache, as no objects will be loaded prior to this command running.
I realize that I can execute arbitrary SQL using Session, however, I have the additional requirement that this UPDATE is compatible with the DB it is running on.
So, I want to use TopLink to generate an UPDATE statment perhaps using the query framework. Is this possible? Is there any way to use TopLink's ability to create SQL before passing this into a Session for execution?

In the current versions your only option is to provide the UPDATE SQL directly using the executeCall(new SQLCall("UPDATE ...")).
In 10.1.3 there exists an UpdateAllQuery intended to do such batch update calls. It will additionally invalidate any cached objects if they were to exist.
Doug

UpdateAllQuery

Similar Messages

Maybe you are looking for