Handling row errors

When a pipeline executes, it may be unable to write particular
rows to the destination table. For instance, this could happen with
a row that has the same primary key as a row already in the destination
table.

Using the pipeline-error DataWindow

To help you handle such error rows, the pipeline places them
in the DataWindow control you painted in your window and specified
in the Start function. It does this by automatically
associating its own special DataWindow object (the PowerBuilder pipeline-error
DataWindow) with your DataWindow control.

Consider what happens in the order entry application. When
a pipeline executes in the w_sales_extract window,
the Start function places all error rows in the dw_pipe_errors DataWindow
control. It includes an error message column to identify the problem
with each row:

Making the error messages shorter If the pipeline’s destination Transaction object
points to an ODBC data source, you can set its DBParm MsgTerse parameter
to make the error messages in the DataWindow shorter. Specifically,
if you type:

then the SQLSTATE error number does not display.
For more information on the MsgTerse DBParm,
see the online Help.

Deciding what to do with error rows

Once there are error rows in your DataWindow control, you
need to decide what to do with them. Your alternatives include:

• Repairing some
  or all of those rows
• Abandoning some or all of those
  rows

Repairing error rows

In many situations it is appropriate to try fixing error rows
so that they can be applied to the destination table. Making these
fixes typically involves modifying one or more of their column values
so that the destination table will accept them. You can do this
in a couple of different ways:

• By letting the user edit one
  or more of the rows in the error DataWindow control (the easy way
  for you, because it does not require any coding work)
• By executing script code in
  your application that edits one or more of the rows in the error
  DataWindow control for the user

In either case, the next step is to apply the modified rows
from this DataWindow control to the destination table.

To apply row repairs to the destination table:

1. Code
   the Repair function in an appropriate script.
   In this function, specify the Transaction object for the destination
   database.

2. Test the result of the Repair function.

For more information on coding the Repair function,
see the PowerScript Reference
.

Example

In the following example, users can edit the contents of the dw_pipe_errors DataWindow
control to fix error rows that appear. They can then apply those modified
rows to the destination table.

Providing a CommandButton When painting the w_sales_extract window, include
a CommandButton control named cb_applyfixes.
Then write code in a few of the application’s scripts to
enable this CommandButton when dw_pipe_errors contains
error rows and to disable it when no error rows appear.

Calling the Repair function Next write a script for the Clicked event of cb_applyfixes.
This script calls the Repair function and tests
whether or not it worked properly:

Together, these features let a user of the application click
the cb_applyfixes CommandButton to
try updating the destination table with one or more corrected rows
from dw_pipe_errors.

Canceling row repairs

Earlier in this chapter you learned how to let users (or the
application itself) stop writing rows to the destination table during
the initial execution of a pipeline. If appropriate, you can use
the same technique while row repairs are being applied.

For details, see “Canceling pipeline
execution”.

Committing row repairs

The Repair function commits (or rolls back)
database updates in the same way the Start function
does.

For details, see “Committing updates
to the database”.

Handling rows that still are not repaired

Sometimes after the Repair function has
executed, there may still be error rows left in the error DataWindow
control. This may be because these rows:

• Were modified by the user or application but still
  have errors
• Were not modified by the user or application
• Were never written to the destination table because
  the Cancel function was called (or were rolled
  back from the destination table following the cancellation)

At this point, the user or application can try again to modify
these rows and then apply them to the destination table with the Repair function.
There is also the alternative of abandoning one or more of these
rows. You will learn about that technique next.

Abandoning error rows

In some cases, you may want to enable users or your application
to completely discard one or more error rows from the error DataWindow
control. This can be useful for dealing with error rows that it
is not desirable to repair.

Table 17-4 shows
some techniques you can use for abandoning such error rows.

Table 17-4: Abandoning error rows

If you want to abandon	Use
All error rows in the error DataWindow control	The Reset function
One or more particular error rows in the error DataWindow control	The RowsDiscard function

For more information on coding these functions,
see the PowerScript Reference
.

Example

In the following example, users can choose to abandon all
error rows in the dw_pipe_errors DataWindow
control.

Providing a CommandButton When painting the w_sales_extract window, include
a CommandButton control named cb_forgofixes.
Write code in a few of the application’s scripts to enable
this CommandButton when dw_pipe_errors contains
error rows and to disable it when no error rows appear.

Calling the Reset function Next write a script for the Clicked event of cb_forgofixes.
This script calls the Reset function:

Together, these features let a user of the application click
the cb_forgofixes CommandButton to
discard all error rows from dw_pipe_errors.

---