Merge changes back to the client

Save PDF

Last Updated: January 22, 2026
2 minute read

OpenEdge
Version 12.8
Documentation

When the server returns the change dataset back to the client, the returned dataset may not match the working dataset of the client because the dataset may contain changes made by database triggers or by other users. These server-side changes must be merged into the working dataset on the client.

Merging changes after an update is different from retrieving the data. When you retrieve the data, you do not know what changed. When you merge changes, you can examine every record to process the changes or errors that occurred as a result of the update call.

You write code to merge the changes returned from the server and clear the before-table in the working dataset. In some cases, the returned dataset may contain errors.

Merge changes back to the working dataset

You use the MERGE-CHANGES( ) method to merge any server-side changes back into the working data in the client. When you merge the changes, the before-table for the working dataset is cleared. Before you call MERGE-CHANGES for a returned change dataset, you should ensure that there are no errors in the returned change dataset, by checking the ERROR attribute.

In the following example code, these steps are performed on the client:

The client sends the change dataset to the server for update.
The change dataset is returned from the server.
If no errors are found, the changes are merged back into the working dataset and the before-table of the working dataset is cleared.

/* client.p */
. . .
/* update the data in the server */
RUN updateOrderOrderLineDS IN hProc(INPUT-OUTPUT DATASET-HANDLE hChangeDataSet).
IF hChangeDataset:ERROR = FALSE
THEN DO:
    hChangeDataset:MERGE-CHANGES(DATASET dsOrderOrderLine:HANDLE).
END.

Datasets