AppletTalk.com

server · Guest

message unavailable

Thomas Kellerer · Guest

On 06.02.2006 12:00 snehalapatil (AT) gmail (DOT) com wrote:

Robert Klemme · Guest

Thomas Kellerer wrote:

Joe Weinstein · Guest

snehalapatil (AT) gmail (DOT) com wrote:

Guest

Yes, I did try ,

CallableStatement s = c.prepareCall("{call spGetGrid(?,?)}");
s.setString(1,"PRJ-001-A");
s.setClob(2, null);
s.registerOutParameter(2, java.sql.Types.CLOB);
ResultSet r = s.executeQuery();

but now im getting the error message:
Implicit conversion from data type xml to nvarchar(max) is not allowed.
use convert function to run this query.

I am not sure what has to be done to convert this output type (xml) to
clob.

Any help would be greatly appreciated.
Nathan

Green · Guest

What database and driver are you using?

David Harper · Guest

Mradul wrote:

snehalapatil@gmail.com · Guest

using sybase database

Green · Guest

Besides other suggestions, you want to look into your JDBC connection
properties. Depending on your driver, there are 2 properties which
typically determine how resultsets are cached and returned to client.
They are:

1. InsensitiveResultSetBufferSize
Determines the amount of memory used by the driver to cache insensitive
result set data.

2. SelectMethod
A hint to the driver that determines whether the driver requests a
database cursor for Select statements. Performance and behavior of the
driver are affected by this property.

Both are optional and the default for SelectMethod does not work well
for large resultsets. Check your driver documentation about this.

joeNOSPAM@BEA.com · Guest

Are you using Sybase's driver (jconn2.jar or jconn3.jar?).
Can you tell us why you're retrieving 79,000 rows?
Sybase gives plenty of capability in their stored procedures
to allow you to avoid that.
Joe

snehalapatil@gmail.com · Guest

using jconn2.jar
& have to retrive retrieve 79,000 rows without using stored procedures
-snehal

joeNOSPAM@BEA.com · Guest

If you absolutely have to retrieve 79,000 rows, are you processing them
one-by-one, or do you have to collect them all in an array/list of
some sort?
The only thing you can do is to select only those columns you need,
and id you can process the rows one-by-one, look at the jConnect driver
docs about memory management. You should be able to specify your
result sets to use server-side cursors and read-forward-only so the
driver
itself doesn't have to cache all the data at once. Does the table have
a
unique index? If so, you can write fancier SQL to select the first 1000
rows,
process them and then re-use the memory querying for the next 1000 till
you're done.
If I were you I would at least get some intelligence points by
mentioning
to my higher-ups that whatever processing these 79000 rows need would
be
much better done in stored procedures in the DBMS where the data is.
Good luck...
Joe Weinstein at BEA Systems

Mradul · Guest

Hi David,

Thanks for the solution, this could be the case. Coz when i'm pinging
my system, it is showing the IP address assigned to me by my ISP.

Thanks again

Regards,

Mradul Kaushik

David Harper · Guest

Mradul wrote:

Jan Dieckmann · Guest

Spring (www.springframework.com) offers excellent support. Also look at
Hibernate. Hibernate actually is an ORM tool.

regards Jan Dieckmann

"GG" <GG (AT) gg (DOT) pl> schrieb im Newsbeitrag
news:dqg6s2$gj0$1 (AT) nemesis (DOT) news.tpi.pl...