DSXchange

1. Pre-load a hashed file with the CURRENT records from the target, with the business, or natural, key as the key column, the surrogate key as a non-key column, and any columns needed for change detection also as non-key columns. 2. In a Transformer stage check for existence. If not found, generate ...

You CAN have RCP. Design it in Designer, compile it and see what develops.

Some people have a life on weekends.

C'est moi.

Why not Search for it?

I don't believe in causing jobs to abort. I feel it is far better to retain control over what is happening. That said, there are several solutions, including limiting the number of rows on an output link from a Transformer stage that might handle this particular value, or conditionally calling a rou...

Technically, because there are no data types in a text file (only text), text files do NOT support NULL. Therefore your metadata need to describe how NULL is represented within the file/field.

Ereplace() would be the more appropriate function, since Convert() only performs a character-by-character conversion. The Convert() solution would also eliminate isolated LF and CR characters, possible incorrectly.

Vertical pivot is allegedly to be available in the PXPivot stage (subject to the usual caveats) in the next version (probably numbered 8.5, subject to other caveats).

Disable operator combination and try again, to determine in which stage this is occurring. At that point we may be able to offer some cogent advice.

Tell is what "didn't work" actually means. Usually this is implemented either through a job parameter or through a reference link to which the returned value has @INROWNUM added in a Transformer stage. However, this involves doing the query for every row processed, which is why performing ...

The profiling can be automated (through the iaJob command).

Really, this process is too complex to spend time on here. Involve your official support provider. We don't have the necessary diagnostic tools. You may, for example, have missed a vital patch after FP1.

Please wrap your design in Code tags and use Preview till it appears comprehensible. I have no idea what the unformatted design is supposed to be doing.

This is what's called an "unbounded VarChar".

These have a specific storage strategy in Data Sets, which ordinarily store data (even bounded VarChar) in fixed width format, because it's far more efficient to pack and unpack.