DSXchange

.

sorry(And Thanks) craig,

Note to self: Do not post just after getting up!!

Read up on the Pivot Stage in your Ascential Documentation.

you can set the field to non-nullable.


I'd recommended the nullable option becuse of:

How can I get it to insert the word NULL if the lookup fails !

in the PX Funnel stage, all the input datasets must have exactly the same metadata as each other and the output dataset. There can be no difference in the number, data type and (i think) precision and nullability. Check every input link into that funnel stage. It should have the same column definiti...

Ok, In that case, undo this step i had mentioned: you will also need to set the nullability of this field to "Yes" and switch RCP off. In a transformer, set the derivation of the necessary output column(new or existing) to: if in.col="" then 2 else 1 in.col is the lookup referenc...

Hi Aliso,

Need a clarification:

What output do you need if the lookup fails. Is it "2" or the null value or the string "NULL" ?

If you are using Version 7.5, you don't need the nested condition stage. The routine stage supports conditional triggers based on returnvalue.

On lookup failure, the default value for an non nullable varchar field is empty string "". I don't remember if the modify stage has a function for converting empty strings, but a transformer will convert this to a null. the derivation for this would be: if in.col="" then setNull(...

If I remember right (No access to PX now ), select the "User-defined Update Only." mode and write an insert statement instead.

It works.

What I meant was: In the transformer (I assume) that splits the single record(I assume this to be row of data) into Policy Details and Address Details, add another column to both output streams; set derivation of both to @INROWNUM. this new column can be used as the hash key. Sorry if i was not clea...

AFAIK the only 2 operations supported by sequential files are append and overwrite. The sequential file stage has no mechanism for updating previously written rows. if it is imperative that output be a text file, you will need to design logic that re-creates the entire file. One thing that comes to ...

Recommend using the row number(@INROWNUM) as the hash key.

When DataStage EE finds that a target columns metadata does not match the source's <B>and</b> no explicit conversion has been made but an automatic conversion is possible( like converting from int16 to int32), it does the conversion and warns(informs?) the operator that DataStage had to do an operat...

Hi Raj9176, Please post server job posts to the Server Forum. For your issue, there are 2 solutions: 1. Pure DataStage: (Inter process row buffering needs tobe switched on for this) Generate header, detail and trailer information in 3 different streams in a single job. In the final transformer for e...

Just thinking out loud:

Cant the function be called in the target stage in the insert( and update?) statement(using user-defined sql) rather than doing a lookup in DataStage?.


Are there any manipulations being done in DataStage on the value returned by the UDF?