DSXchange

This is one of suggestion. As you get daily data, you must process it daily to complete it fast. You can process the 10 million data, aggregate it daily and do summarized operations. At the end of the month you can again do a summary or rollup of all the daily summarized data. At this time you will ...

It depends on the way you are going to use it again. Mostly error tables are checked rarely. You are inserting with error number,error code. So why seperate table, store it in single table.

Bury all errors in single table.

Its because you need to FTP files from DOS to UNIX only in ASCII format , so that host transfers data consistently. If you have already transfered then use "dos2unix" command for converting DOS files to UNIX files. Also while reading Fixed width make sure you are assigning exact column siz...

As you can't directly call a sql statement in before job subroutine. You can write a small unix script to do that and call the unix script through Execute SH command.

Else write a small procedure and call the same by ODBC stage before calling the actual stage

I think you requirement is to create a customized Config file for a particular job. You have specified you job type as server. In server you dont have the concept of Configration files. If its parallel you can create as may config files you need. Just copy you original config file and backup it. The...

Yes. If my understanding is right. You need licence for only either Server or Parallel. If dont want parallel jobs , then you can get only server job working from IBM. In our last project we got only server working.Though you can see parallel options, they will be disabled.

Is that you need?

Did you try using Double quotes. Else pass it in a variable and then use it by calling it.

I have also faced this problem.
Check if the file count is matching the number of inserted records.

A.May be it, contains some header or footer.
B. Check if your job is no overwriting the error count setting.

Check if any after job soubroutine is being called in the job, which is making it to fail.

I dont think you will save any time by converting the sqlldr to OCI stage. Because the time spent in moving the data to Flat file to Dataset must also taken in to count. so some the options are 1. Increase your parallelism to 4 then you can see some change. We dont your server config . so try by inc...

Did you try executing the same from any of the Oracle Client. If it is working then try the same in Datastage. Try to pass the value of the parameter as including double quotes.

But this is not the standard table naming convention.

For example you have a script as Check_File.sh then you can call it as follows. 1. Select ExecuteSh in AfterJobSubroutine and enter the following <Script_Path>/<Script_Name>.sh #Parameter1# #parameter2# The parameter should be declared in parameter list in job. Example: /u10/files/script/Check_File....

Please do not post the same question in different forums. Please delete it.

You must redirect the file output incase you opt to use sed or awk. Rather try to use perl command to find & replace in the same file without redirecting. Incase if the file is large, perl -pi will be much faster than sed. For example try perl -pi -e 's/<old text>/<new text>/g;' Other Option As ...

Try if (InputColumn) = '' then 01/01/1800 From your post it is clear that your input is timestamp or blank value, but what is your input datatype , is it Timestamp or string. what is your output datatype Date or Timestamp. Accordingly you can convert it to TimeStampToDate or DateToTimeStamp. But Why...

Mostly this type of errors occur when you try to convert from one datatype to other(INT to DECIMAL or FLOAT) . Check if conversion is from same datatype to datatype.
Check if incoming data is correct value for the datatype.
Check for size and precision.