Oracle-specific data types vs Connection pool

Hi, using Oracle-specific data types such as
ArrayDescriptor and create it by a connection to an Oracle database from a WebSphere datasource throw a ClassCastException. There is no problem using these Oracle-specific data types when the connection is obtained via the DriverManager. WebSphere cannot fix the problem because these data types are
not part of the JDBC specification.
Some reply from Oracle?
Thank you very much.

Thank you but I there is no solution in that thread.
Also, I know that WebLogic from BEA has a way around this issue... No Websphere work around ?
No Oracle fixing ?
Thanks.

Similar Messages

ADF View Object query to use Oracle LONG data type

Hi,
I'm using a view object that uses Oracle LONG data type.The following exception is thrown
*"java.sql.SQLException: Stream has already been closed"* and it is not fetching the correct data.
I have searched in many forums,they suggested to use LOB instead of LONG as LONG is deprecated. But in my case I can't avoid using LONG as some of my legacy applications using the same DB.
Please suggest me a way
Thanks in Advance
Felix

Thanks for your reply Arun,
I have solved this problem using the solution suggested by this blog [AskTom-Link|http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:839298816582]
In the above blog link,they have suggested to convert Long to Varchar by functions and asked to use this function in the column.
Thank you all guys who have seen this post.
Regards,
Felix

Enforcing integrity using oracle specific data base commands .v. using fron

Full subject : Enforcing integrity using oracle specific data base commands .v. using front end.
It ought to be generally accepted that it is better to enforce integrity directly in the data base using constraints, dependencies, triggers etc rather than leaving it to specific front end programmes.
In my view, the chief advantages - of enforcing integrity directly in the data base - are:
(1) No process can violate the integrity.
(2) Being server centric, these commands run on the server and so can be easily sized in one place.
(3) One time data migration (imports) into the system using Oracle tools or SQL commands would also need to conform to the integrity constraints. Thus an implementor would be assured that the basic data is alright.
I am faced with a situation where we are about to implement a new Oracle based package. During data migration, when we discovered that there are no integrity constraints built into the target data base, the package vendor asserted that it is not necessary to build in integrity into the database. This seems to be an extreme and risky view.
Further, it is argued by the package vendor that putting constraints directly in the data base would significantly increase the needed resources (RAM) on the server. In my view, this increase is trivial and anyway, hardware costs are crashing day by day.
In the absence of integrity checks in the data base, it seems to me that every program would have to extra zero value work to ensure integrity of the end user data. It will still never be complete.
I would like to know the pros and cons of implementing without integrity constraints.
OK.

AnanthaP wrote:
Full subject : Enforcing integrity using oracle specific data base commands .v. using front end.
It ought to be generally accepted that it is better to enforce integrity directly in the data base using constraints, dependencies, triggers etc rather than leaving it to specific front end programmes.
In my view, the chief advantages - of enforcing integrity directly in the data base - are:
(1) No process can violate the integrity.
(2) Being server centric, these commands run on the server and so can be easily sized in one place.
(3) One time data migration (imports) into the system using Oracle tools or SQL commands would also need to conform to the integrity constraints. Thus an implementor would be assured that the basic data is alright.
I am faced with a situation where we are about to implement a new Oracle based package. During data migration, when we discovered that there are no integrity constraints built into the target data base, the package vendor asserted that it is not necessary to build in integrity into the database. This seems to be an extreme and risky view.
Further, it is argued by the package vendor that putting constraints directly in the data base would significantly increase the needed resources (RAM) on the server. In my view, this increase is trivial and anyway, hardware costs are crashing day by day.
In the absence of integrity checks in the data base, it seems to me that every program would have to extra zero value work to ensure integrity of the end user data. It will still never be complete.
I would like to know the pros and cons of implementing without integrity constraints.
OK.It's a shame you seem to be so far into the process and committed to this vendor. I once had a vendor tell us his product would run on Oracle but they recommended MS SQL Server because "oracle can't handle more than 5 concurrent sessions." I made sure that vendor didn't make the short list.

Problem loading PostgreSQL Bytea data type to Oracle Raw data type

We are migrating our database from PostgreSQL to Oracle. First, we convert the BYTEA data type in PostgreSQL to Oracle RAW. The BYTEA data type is variable bytes array. How can we load the BYTEA data type to Oracle RAW data type? Or I have to convert to different data type. thanks.
Peter,

hi,
Instead of 'interval day to second' in method declaration use internal datatype 'DSINTERVAL_UNCONSTRAINED'.
There are more unconstrained types in oracle.
Bartek

Changing Data source in connection pools

hi all,
When the primary data source goes down or is unresponsive, we have to often point our repository connection pool to a backup data source. As this is production environment, we would like to do this without bringing the services down.
I usually make sure that there are no requests running by going to Manage Sessions and clicking on 'close all cursors' and 'Cancel all running requests'. Are there any other implications of not restarting the services while changing the data source connections? Do the connection pools are kept open to the old data source?
We use Oracle 10G as a database, but I have also done this with DB2 database.
Thanks in Advance

Turribeach said it correct.
But if you are handling the failover manually for now until you figure out the RAC setup, it is better you kill all the user session through the admin tool . AdminTool -> Manage -> Sessions
Otherwise the users already connected will keep getting errors until they choose to relogin or choose to "Reload Server Metadata" in the answers, which they would probably not even know. So it is better to kill their session forcing them to relogin. But the transition will not be seamless if they in the process of firing a query.

Can't update oracle Long data type

Hi All,
Setup
Weblogic 5.1.0sp11, oracle 8.1.6, Solaris8, classes12.zip,
Solaris_JDK_1.2.1_03a
ConnectionPool:
weblogic.jdbc.connectionPool.XXXDBPool=\
url=jdbc:oracle:thin:@10.2.30.50:1521:ZZZZ,\
driver=oracle.jdbc.driver.OracleDriver,\...etc
Problem
I can't seem to write to (insert/update) an oracle type Long column
using connections from the weblogic connection pool. Each entry(as
XML) into this column is fairly large (>1MB). Reading from works fine,
however.
The same configuration(minus connection pool and associated
code/config) via tomcat works fine(read/write), so I'm thinking the
weblogic connection pool is giving me a connection object which
doesn't update this Long column properly.
Using the usual(again works fine for tomcat) :-
//Url = jdbc:weblogic:pool:XXXDBPool
//driver = weblogic.jdbc.pool.Driver
stat = con.prepareStatement("UPDATE BM_RM_USER_DIARY set DIARY=? where
id=?");
//stat.setString(1,diaryData); //use an inputstream instead.
ByteArrayInputStream bais = new
ByteArrayInputStream(diaryData.getBytes());
//StringBufferInputStream bais = new
StringBufferInputStream(diaryData); //deprecated
stat.setAsciiStream(1, bais, bais.available());
stat.setLong(2,userId);
I'm not getting any errors from this update, so I can't really see
what I might be doing wrong. It just returns fine as if it has
succeeded (rows updated = 1)
Assuming I CAN NOT change these settings, are there any other ways I
can programmatically get these inserts working ? Is this even a known
issue ? as i can't find a problem in these postings that quite matches
mine.
Any help appreciated.
Cheers
Alkesh

Hi All
Just to follow up on this.
The problem was solved by explicitly calling commit() after
the update has taken place.
It seems as though my connection from the connection pool has
autocommit set to false.
Thanks
Alkesh
[email protected] (Alkesh) wrote in message news:<[email protected]>...
Hi All,
Setup
Weblogic 5.1.0sp11, oracle 8.1.6, Solaris8, classes12.zip,
Solaris_JDK_1.2.1_03a
ConnectionPool:
weblogic.jdbc.connectionPool.XXXDBPool=\
url=jdbc:oracle:thin:@10.2.30.50:1521:ZZZZ,\
driver=oracle.jdbc.driver.OracleDriver,\...etc
Problem
I can't seem to write to (insert/update) an oracle type Long column
using connections from the weblogic connection pool. Each entry(as
XML) into this column is fairly large (>1MB). Reading from works fine,
however.
The same configuration(minus connection pool and associated
code/config) via tomcat works fine(read/write), so I'm thinking the
weblogic connection pool is giving me a connection object which
doesn't update this Long column properly.
Using the usual(again works fine for tomcat) :-
//Url = jdbc:weblogic:pool:XXXDBPool
//driver = weblogic.jdbc.pool.Driver
stat = con.prepareStatement("UPDATE BM_RM_USER_DIARY set DIARY=? where
id=?");
//stat.setString(1,diaryData); //use an inputstream instead.
ByteArrayInputStream bais = new
ByteArrayInputStream(diaryData.getBytes());
//StringBufferInputStream bais = new
StringBufferInputStream(diaryData); //deprecated
stat.setAsciiStream(1, bais, bais.available());
stat.setLong(2,userId);
I'm not getting any errors from this update, so I can't really see
what I might be doing wrong. It just returns fine as if it has
succeeded (rows updated = 1)
Assuming I CAN NOT change these settings, are there any other ways I
can programmatically get these inserts working ? Is this even a known
issue ? as i can't find a problem in these postings that quite matches
mine.
Any help appreciated.
Cheers
Alkesh

How could I Write data into a field in Oracle whose data type is VARCHAR2

The target data I want to write into Oracle is in http://tw.yahoo.com/info/utos.html.
Now, these data is stored in Mysql database and the field which stores these data uses "Text" as its data type.
I want to derive these data from mysql database and store them into a field of oracle database.
In oracle, I create field whose data type is varchar2(4000) to store these data.
I use JSP to derive data from mysql and insert into oracle through JDBC.
But the result of the page shows me that "javax.servlet.ServletException: Data size bigger than max size for this type: 25623".
Please anyone could help to resolve this problem?
Thank you very much.

My hunch is that the problem is that VARCHAR2(4000), but default, allocates 4000 bytes of storage. Depending on your database character set, a single character may require up to 4 bytes of storage.
If you are on 9i, you can declare the column VARCHAR2(4000 CHAR) to allocate 4000 characters of storage. You can also set NLS_LENGTH_SEMANTICS to CHAR, which will cause Oracle to assume that your declarations are in characters rather than bytes.
Justin
Distributed Database Consulting, Inc.
http://www.ddbcinc.com/askDDBC

Mapping Data Source and Connection Pools

Is it possible to Map a Data Dource with Multiple connection pools in Weblogic ? If Yes how do we do that ?

Which version of WLS are you using. With version 10.0 you could use a multi data source.
http://e-docs.bea.com/wls/docs100/jdbc_admin/jdbc_multidatasources.html

Data type in module pool

I want to create sceen in module pool programming with 2 filed , one with type P length 8 decimals 2, and another with type I,
as I need it to do calculation.
can anybody suggest ?
thanks

Hi,
Its depend on your requirements-
If you want To display text then go for CHAR Type.
If you want to display number then Go for INT4/INT2/INT2
For amount DEE - So On.
Now Put the Text filed from List And Double Clicking on the field Attribute list is diaplyed , In the Distionary Attributes you put DATA TYPE.
Do not Forgrt to place OK-CODE Field In Element List.In the Element list you also Define the Type, Length attrubutes.
If Any Confusion let me know.
Regards,
Sujit

ORACLE Parameter Data Types

Hi,
We are using for all Oracle number columns the Oracle parameter type OracleType.Number.
Now we have seen that we can also use the type OracleTyp.Int32.
Is it better or faster to use this type for columns
which can use this type.
Have anybody experience with using Oracle data types ??
Kind Regards
Martin

It really depends on what kind of data you're inserting or updating. If you've data that can be accomodated in an integer, you should use OracleDbType.Int32 over OracleDbType.Decimal. You get much better performance by using Byte, Int16, Int32, Single, Double over Decimal but these types can not hold the same precision and scale as Decimal type can.

How to avoid oracle implicit data type conversion

Hello,
i have a query like below..
SELECT EMP_NUM
FROM EMPLOYEE
WHERE EMP_NUM=SOURCE_EMP_NUM;
1. My SOURCE_EMP_NUM data type is varchar2.
2. My table EMP_NUM data type is number.
3. In one case my SOURCE_EMP_NUM value '00000123'
4. I have an entry in table for EMP_NUM as 123.
5. In this case my query is success and giving me row for 123, which i don't want it should compare with leading zeros. But due to oracle implicit conversion the comparision getting passed.
6. I have an index for this column, so i can't use "like" operator...
Can somebody help me.....
Thanks in adv....

HI,
Just see this :---
SQL> select * from emp where deptno=00020;---------deptno is number, 00020 is number and its true
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPTNO
7369 SMITH CLERK 7902 17-DEC-80 800 20
7566 JONES MANAGER 7839 02-APR-81 2975 20
7788 SCOTT ANALYST 7566 19-APR-87 3000 20
7876 ADAMS CLERK 7788 23-MAY-87 1100 20
7902 FORD ANALYST 7566 03-DEC-81 3000 20
SQL> select * from emp where to_char(deptno)=00020;=-------------deptno is character type and 00020 is number so still it is leaving 0000 from 20 as it number and has no impact,
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPTNO
7369 SMITH CLERK 7902 17-DEC-80 800 20
7566 JONES MANAGER 7839 02-APR-81 2975 20
7788 SCOTT ANALYST 7566 19-APR-87 3000 20
7876 ADAMS CLERK 7788 23-MAY-87 1100 20
7902 FORD ANALYST 7566 03-DEC-81 3000 20
SQL> select * from emp where to_char(deptno)='00020';---now both are character and no row is selected.
no rows selected
So you can avoid the implicit conversion according to the requirements and in this case we are usin to_char function to avoid this.
Regards,
Navneet

Oracle Object Data Type and EJB

Hi, All:
Has anyone implemented a project with Oracle user-defined data type (Object
Data Type) together
with EJB? That is, using ORDBMS (Object Relational Database) together with
EJB. I am thinking to use ORDBMS together with EJB. Any information, samples
and references to this direction is appreciated.
Thanks,
John Wen
RM2M, LLC
[email protected]

I''ve got a solution! I've had sql in my query
SELECT
ltrim(OPERATIVE_REPORT."orderText",11),
OPERATIVE_REPORT."resolID",
COUNT(OPERATIVE_REPORT."(usrID)order_executionUser")
FROM
OPERATIVE_REPORT
GROUP BY
ltrim(OPERATIVE_REPORT."orderText",11),
OPERATIVE_REPORT."resolID"
where orderText has clob datatype. And the reason is GROUP BY. GROUP BY deny CLOB usage.

Oracle apps data type number came over to disco as varchar

created custom foler ( fnd_flex_values_vl) using sql, one of DFF item ( original data type is number) shows as varchar. Anybody knows how to convert this to number , I tried to_number but still shows as text in Desktop.

Sorry, thought you'd said it was a function originally.
Okay, so as a custom folder, is it possible that there are NULLs being encountered in the attribute1 field?
ie:
select flex_value_set_id, flex_value_id,flex_value, description,
to_char(attribute1,'999,999,999') budget,to_number(attribute1,'999,999.99') Budget_amount_number
from fnd_flex_values_vl
where flex_value_set_id='1004012'
becomes:
select flex_value_set_id, flex_value_id,flex_value, description,
to_char(NVL(attribute1,'0'),'999,999,999') budget,to_number(NVL(attribute1,'0'),'999,999.99') Budget_amount_number
from fnd_flex_values_vl
where flex_value_set_id='1004012'
or is it something more difficult?
Russ

ORACLE CLOB data type and WIS 10901

Hello! I have several objects in universe that has connection to ORACLE DB. One of them has type CLOB in db and Long Text in universe, other has type number and has definition COUNT(db_column).
When I'm using this objects separate in webi query - everything fine, but when I put them together I've got error message:
A database error occured. The database error text is: ORA-00932: inconsistent datatypes: expected - , got CLOB
. (WIS 10901)
What's going wrong?

I''ve got a solution! I've had sql in my query
SELECT
ltrim(OPERATIVE_REPORT."orderText",11),
OPERATIVE_REPORT."resolID",
COUNT(OPERATIVE_REPORT."(usrID)order_executionUser")
FROM
OPERATIVE_REPORT
GROUP BY
ltrim(OPERATIVE_REPORT."orderText",11),
OPERATIVE_REPORT."resolID"
where orderText has clob datatype. And the reason is GROUP BY. GROUP BY deny CLOB usage.

Oracle Long Data type Problem

Hi,
I am using ORACLE 7.3.4 database. I am having a problem in updating a column of long datatype in a table t1 with another column from table t2 of long datatype. I have tried setting LONG 1000000 also but this doesn't work. I appreciate any kind of suggestions.
The error it gives is ORA-00997: illegal use of LONG datatype.
Thanks
Reddy

I had the same problem too with a newest thin driver. I found that I can read all my characters (bytes) using getBinaryStream() and single-byte read(), but read() never returns -1; instead reading the character before last causes the ArrayOutOfBoundsException you mentioned.
So I just blindly caught the exception and let it go to QA. (sort of
InputStream r = rs.getBinaryStream(field);
if(r == null) ...
StringBuffer sb = new StringBuffer();
try { for(int c = -1; (c = r.read()) != -1;) sb.append((char) c); }
catch(IndexOutOfBoundsException e)
{} // ignore Oracle idiosyncrazy.
finally { r.close(); }
). Please note that getCharacterStream() or getAsciiStream() did not work for me. I did not try deprecated getUnicodeStream(): I decided I would better correctly extract characters in Java from binary data myself if I need them rather than allowing this junky thin Oracle driver to do an extra thing for me.

Oracle-specific data types vs Connection pool

Similar Messages

Maybe you are looking for