DSXchange

In short, if you take out the first 37 records from input, will the resulting flow work lead to full load?

Steps B and D are identical but yet said to result in different values.

Are you using any stage variables and derivations in them?

Can you list the derivation of the target column tracing from its source. Something like Craig asked.

What if you get the link from very first transformer without any derivation ?

Craig,

Being Server edition, I did think about deleting or including a constraint to skip 37.

But that is on the assumption that the target is direct load of source with no other dml or "prev-next" included.

Are you saying that it had happened or you want to prevent such happening? Unless there is any metadata related information to locate records affected by current load, you cannot restart from failure. Alternatively you can set the commit size to zero to ensure that it is everything or nothing. There...

Removing trailing spaces is default property for Varchar.

Branch the result immediately after Oracle source and check the result.

Priyadarshi, Using ODBC drivers for text files may be simpler option. Creating external tables in Oracle is not straightforward. You have several hurdles to clear - such as create oracle directory, grant priv to the user, create external table ddl, grant select to the user. Also this must be visible...

Try commenting doTransfer.

Also check default properties for data transfer.

What is the source of the 'missing' column?

If it is SQL, can you post the SQL here?

If it is from DataSet, can you check whether it has the mentioned column?

Trace it back from the modify. You appear to be missing the mentioned column.

Is the source from user-defined SQL?

Create a seq file as output from the job with your condition and output as "del A.txt" or "del B.txt"

In the post-job, execute the resulting script file.

What do you wish to achieve? How will you control which part of the dataset you will use in a job?

If it is just the size of partitions, you can copy the dataset using orchadmin into different configuration. But that is not straight-forward.

D U P L I C A T E