DSXchange

I have a job as follows:

DataSet -> Xfm1 -> Modify -> SequentialFile

I have 1 specification in the Modify stage as:

OutCol1:decimal[10]=decimal_from_string(InCol1)


The job runs fine other than a few warnings. I had questions about 2 specific
warnings:

1)Xfm1: When checking operator: Dropping component "OutCol1" because of a prior component with the same name.

2)Xfm1: When checking operator: When binding output schema variable "APT_TRoutput0Rec0": When binding output interface field "InCol1"
to field "OutCol1": Conversion from source type "string[12]" to result type "decimal[10,0]": Possible range limitation.

This doesn't make sense to me. My questions are:

1)On warning #1:
There is no column in either the input or output tab of the transformer named OutCol1.

2)On warning #2: Why would these warnings come from the transformer stage? There is no column in either the input or output tab of the transformer named OutCol1. This column is on the output tab of the modify stage. There is no conversion happening on the transformer. It is being used as a straight copy.

I can't understand that why do you use a transformer in between DataSet and Modify Stage. Modify stage is a light weight Transformer. Can you drop the Transformer stage and link the output of DataSet to Modify Stage. This may resolve your problem.

Un-check your RCP, then warning messages will be gone.

I added the transformer because I didn't want to use a KEEP specification for so many columns. Anyway, I removed it and I also unchecked RCP. Now, I get the warning on the target sequential file stage:

SequentialFile: When checking operator: When binding input interface field "LoanNumber" to field "LoanNumber": Implicit conversion from source type "decimal[10,0]" to result type "int32": Possible range limitation.

This may be due to different types of Metadata defined for LoanNumber column. Make sure that both input/output has same type of metadata. In this case specially precision should be same for both input/output.

This may be due to different types of Metadata defined for LoanNumber column. Make sure that both input/output has same type of metadata. In this case specially precision should be same for both input/output.

Its a straight copy from the Modify stage to the SeqFile stage. Actually, whatever is in the output tab of the Modify stage goes automatically to the seqfile stage so there is now way that metadata will be different.

Can you provide the following details.
1. What is the datatype of Input tab of Modify Stage (LoanNumber)
2. What is the datatype of Output tab of Modify Stage (LoanNumber)
3. What is the Specification for this column ?

Input data type is char(12) and output is integer(10). The specification I used is:

OutCol1:decimal[10]=decimal_from_string(InCol1)

A different question to everyone. To convert from string to integer, it seems like I have to do 2 modify stages, one to convert to decimal from string and another to convert to integer from decimal. Does anybody have any other way?

Add one more Specification to the Modify stage, containing the single token NOWARN.

Ray, thank you for your response. I added another specification as:
Specification=NOWARN

This does not suppress the warning on the sequential file. It remains.

Also, do you have an idea about the other question of 2 modify stages?

Since you already have a Transformer stage, you can use nested conversion functions there. Thereby you could avoid any need for your Modify stages.

Yea, I guess for my purposes, transformer is the best. It seems like Modify is much too convoluted to use. As usual, the documentation is terrible. Thanks for your responses.

Have you read the Orchestrate Operators manual? It's more comprehensive.

There are two ways to resolve this issue;

1. Change the length of your input data from 12 to 10

OR

2. Change the specification as follows;
OutCol1:decimal[12]=decimal_from_string(InCol1)
Make sure output column tab also has the same metadata decimal(12,0) for OutCol1.