You are not logged in.
Since you have reported a bug on JIRA, I think you'd better follow MichaŽl Hirt's suggestion.
MS SQL components are optimized for MS SQL JDBC components must remains generic and therefore it might not be optimized for all databases. Any change that we could do here could decrease performance of any other database.
I'm using TIS 4.2.4 under Seven 64bit
Which version of TOS do you use? Which OS, 32bit or 64bit?
I think this is due to your current system environment.
I tested this issue under TOS 5.0.2.
Is performance hit of the same order of magnitude over more rows (i.e. is the problem just that the JDBC connector takes a while to get going?)
Nothing special in tMAP, just pass 2 fields.
I have tested without tMAP and tJDBCOutput as you can found in my screen shot.
Still get bad performances :-(
This is the job I create with instance name and nolock in query.
Is there any special settings in tMap which may have an effect on performance?
I have tested on another job using jtds and simplify it only with Input->tMAP->Output (see JDBCvsMSSQL_1.png)
I still get bad perfomances with JDBC.
I have attached Component property screen shot with URL JDBC, Driver and classeName used.
Do you have any advice for my problem ?
For your information, i'm using TIS 4.2.4. java
Thanks for your support !
I think the bad performance is not caused by JDBC components or MSSQL components.
Because if running them separately, you would get similar performance.
This may be due to your job logic or something else.
Same thing Pedro with 'net.sourceforge.jtds.jdbc.Driver' :
148 rows/s with tJDBC and 6666 rows/s with tMSSql
Any other idea ?
The class name should be 'net.sourceforge.jtds.jdbc.Driver' if you want to use jtds driver.
Thanks for your response Pedro.
But i have tested with jtds driver with following values and get same bad performance:
Please find attached screen capture
Thanks for your support !
In fact, from Java code aspect tJDBCInput and tMSSQLInput are the same.
I have tested them with jtds driver and get silimar performance.
So it may be due to job design.
Or i miss some detail?
i'm facing some very bad performance using tJDBC componants in order to select and load from/into MSSQLserver database.
i don't want to use tMSSQL componants in my case!
For you information i have a job which select top 1000 from table1 and update each line in table1 on one field.
With tJDBCInput only i have 5952 rows/s
with tJDBCInput and tJDBCOutput i have 250 rows/s
with tMSSQLInput and tMSSQLOutput i have 18181 rows/s
JDBC driver is sqljdbc4.jar
Do you have any requirement to improve performance ? Maybe using an another driver ?
I have tested with following drivers also but got same bad performance:
talend_DB_mssqlUtil.jar and jtds-1.2.5.jar
Thanks for your response