DSXchange

$? is the exit status of the final command in the pipeline - in your case the read command. You need different logic to have the shell script exit with the value of the dsjob exit status.

You have been allocated a finite amount of disk space. This is enforced using a quota. You have filled this space. This is not directly a DataStage problem. Clean up any files you no longer need, and ask nicely for the quota to be increased.

You changed your specification. My solution changes only by losing its "- 1". You can use the Fmt() function to enforce leading zero. I remain curious to learn how you would generate these values using ceil without any other arithmetic (since you did not mention any). Switch stage is not available i...

You misunderstand. If you need to convert a string to a date, then you specify the date format of the string (not of the desired date). Indeed, you never specify the format of a date - a date is a binary data type in DataStage parallel jobs. So ask yourself, do you want a date coming out of this, or...

Try processing fewer and/or smaller files, find some solution that does not require a Folder stage (which tries to keep the entire files in memory) - perhaps using a loop in a job sequence, or put more memory in the server.
Search the forum for more ideas.

(From memory) the IEEE standard that prescribes how Float and Double are stored does not permit zero (all zero bits). There is an environment variable that allows you to override this rule but I can't recall its name.

DataStage is not equipped to shared Data Sets between projects. Because you've managed to trick it is no reason to expect that your trick would be supported by Usage Analysis or any other DataStage tool. Probably the best you can do is to export each project and search the export file for the Data S...

Those observations would be just as valid if you used sqlldr without DataStage.
That is, it's got nothing at all to do with DataStage.

It is documented that adjacent Modify operators are not supported. Under what circumstances (full set of Specifications please) are you getting them to compile and execute successfully?

How about you tell us what Specifications you have set up in the Modify and Copy stages? When the score is composed a stage (operator) may be optimized out - you can most easily see this if there is a Copy stage that does nothing. Operator combination can also make it appear that stages "disappear",...

How would you do that with ceil ?!!

Try

Possibly an old DLL hanging around in memory.

The BASIC Transformer stage is deliberately not in the Palette by default, as its use is discouraged (because it has a number of execution limitations, particularly in a multiple machine environment). You can find it in the Stage Types branch of the Repository - and, if you wish, customize your Pale...

A DataStage solution would involve using a job sequence to obtain the field value and construct the file name, which would then be passed to the job that uses it or, alternatively, used to change a generic file name that the job produces.

Less than Rs12 per day, definitely a worthwhile investment in your future.