DSXchange

Welcome Aboard
If you do a search on each and every one of the errors (use key words), you will see that they have been covered in detail before, time and again. Please use the search facility.

Get the three rows, pass it through a transformer where you need to utilize stage variables to do this vertical pivoting. Something to the effect of, compare present key with previous rows key, if same, concatenate the row values with each other. Pass it through remove duplicate stage and retain the...

You can also go for a single sequential file stage as source. make the job multiple instance and run each instance with the three source files one after the other.

Can you have input coming in with a valid scale (00010002.2) ?

Check out Brad's last reply here.. Make sure you converted the COBOL datatypes to DataStage datatypes correctly.
Its one of those decimal fields that seem to be the culprit. Which one, you will have to do some investigative work.

What are you doing in the job? Any sort of arithmatic operation? Maybe you have defined a column as int at getting a value suitable for BigInt. Possible?

Precision is the length of the field and scale is the numbers after the decimal. For exact syntax, check DS help for "Parallel Job Functions" or search here.

It is because your trying to fit a decimal number into a varchar field. It will not work. You need to do StringToDecimal() to fix the lenth and precision and then use DecimalToString() to convert it back to varchar.

How are you accessing the .xls sheet? via odbc or sequential file stage? You need to convert it to csv before reading it via sequential file stage.

Yup. I got it. Thanks once again.

Refer here.

Do an exact search on "has a partitioning method other than entire specified". You will find many posts explaining the root of the warning and how to get rid of it.

How exactly are you running the dsjob command and passing the parameters?
Whatever values need to be determined during run time can be parametrized and its value can be passed during run time.
Another question, how are you running dsjob.exe, used for windows, on a unix box???

Ahh, now I see. For some reason I was stuck on the line terminator. Thanks for the clarification.

In other words, sort by Id and Date, pass it via aggregator, group by Id and Salary, retaining MIN() date, pass it via remove duplicate stage and retain last record (most recent salary).