DSXchange

Method 1: Hack the parameters in the job sequence's status record.

Method 2: Reset or re-compile the job sequence. If you want to use different parameter values, then restartability is invalid.

You can do it with six separate reference inputs to the Lookup stage.

If it's a fixed width format file (as you say) then data types are effectively Char and therefore will not be trimmed.

You have to have the metadata (table definition, record schema) match precisely what is in the file.

Investigate the Folder stage.

Investigate using a Filter command in your Sequential File stage that uses cat to spool all the files into the job as if they were one large data stream.

Please check the target column data type. Indeed, please post that part of the generated OSH that pertains to this stage, with its record schemas, so we can all be sure.

... and the target column is definitely Decimal(4,1) ?

(Don't assure, please check.)

It's the second last tab. You did open the table definition in the Repository, I trust?

You may be surprised how many unwritten business rules you can discover when you profile even the in-house data. Particularly domain-related ones. DBAs are reluctant to add CHECK constraints claiming (correctly) that they slow inserts and updates. Suddenly the business introduces web-based data entr...

This is the fifth time you have posted this question.

Please desist. It wastes space on the DSXchange server.

Good advice has been given in response to some of your earlier posts. Have you heeded it?

You don't need to do anything, provided that the number supplied as a dfloat is small enough to be represented as sfloat. (It is the possibility that it may not be that you are being alerted to.) If you are 100% certain of this, add a message handler to demote the alert to informational.

How about the following?

Even if the rows are selected from a database table, you may need to validate against business rules, if the database is one external to the business (or even internal to the business if data entry is, shall we say, less than robust).

Open a table definition that has a primary key. Go to the layout tab and choose the parallel option. You will find out there what is required. (From memory, since I can't check at the moment, I don't believe there is any mechanism for identifying a key in a record schema, since that item of informat...

Please mark the job type as parallel, not server.

Try converting both the divisor and the dividend to float data type.

I expect that the use of many stage variables (one per field that needs to be checked) is an attempt to achieve maintainability. The svValidRecord (used in constraint expression) is then - logically at least - the AND of all those results. I agree with that approach, purely on the grounds of maintai...