DSXchange

main_program: An NLS map <ISO-8859-1> is specified, but NLS_LANG is not set. The NLS map must be set to an ICU codepage which is equivalent to the character set specified by NLS_LANG segmentation fault Can someone tell me why this problem is occuring.....it occurs many times...sometimes the job wor...

thanks...Also can you suggest what i need to do for the following problem

viewtopic.php?t=119960&start=0&postdays ... ree+memory

Thanks

normally the data is written to 4 nodes...This is through "Auto" partitioning. Is it helpful to do a hash partition on the output dataset in the first job. And then Use the dataset in the second job...and in the transformer stage (next to the input dataset) again set the partitioning as ha...

No...The oracle table is the same source and destination. I have selected columns from the table ....right to a dataset using the 4 nodes in the first job. In the second job i use the dataset and update back to the same table where i selected from. I am already doing same thing what you explained. l...

just to let you know if this matters ...the DS server is at indiana and the database is at southcarolina.

Job1 took 1 hr 20 mins to complete. The source and destination are same.
That is the reason i am writing to dataset fearing deadlocks

Thanks

ArndW wrote:How long does Job 1 take? Combining these two jobs and removing the DataSet will certainly speed up processing.

Hi Sreddy, I have two jobs. First job writes the key columns and the columns i need to update (here 7 key cols + 4 cols need to be transformed and updated) to a dataset. Design: oracle enterprise stage----->transformer----->dataset Second job is used the dataset from the first job and updates the 4 ...

Hi chulett:

What would be the best idea then if i don't get these grants like create and drop table?

Like does it helps splitting the 10 million records into multiple update statements and use each one separately?

Thanks

I don't know your job so I'll make some assumptions. I suppose you have a second source where you have the 7 key colums+ the 4 columns to be updated. If you select all records from your table, all columns exept the 4 and you join this with the second source, you get the full new record. Insert this...

On Oracle I would never update all records of a table. Even if you have 100 columns and only need to update 4 it will be faster to create a new table, load it with the full records(while loading change the 4 columns you need to update offcource.) Once this is done rename the old table(to _archive) ...

yes i am updating each row of 4 columns

Hi, I am updating a table that contains 10 million records. I am updating 4 columns with 7 key columns that make unique for each row. It took 10 hrs to complete 2 million records and if it moves at this pace it takes 50hrs. Can some let me know what is the best approach for it to make it faster? I a...

I am using a c++ routine in the parallel jobs and some tables are huge and this routine is applied on every row of the column. It returns char*. After running some jobs , the next jobs are just getting aborted. I am doing force compile 3 to 4 times and then only the job is running. I don't know if t...

thanks..The dba rectified it...thanks

I am suddenly getting the error ORA-12514: TNS:listener does not currently know of service requested in connect descriptor

Till yesterday everything was fine. This is a Oracle RAC setup.
Suddenly iam getting this error. Can some help me on this?

Thanks