Persistence mode
All source tables have a persistence mode that determines how the new data is processed on the server.
There are three different persistence modes:
OVERWRITE
This is the default setting to follow the standard behavior from previous versions. If this mode is set, the new uploaded data overwrites the existing data. The overwritten data is lost and cannot be recovered. If required, you must upload the overwritten data again.
APPEND (without merge key)
If this mode is set, instead of overwriting the already persisted table data on the server, new uploaded data is appended to the existing data. The new data rows are added at the end in the order in which they are received. This causes the size of the existing source table to grow. Persisting old data a second time with this setting (as if they were new rows) will lead to duplicate entries. This can impact the accuracy of the analysis results.
Note that currently the only way to select this mode for a source table is to use the Data Ingestion API to either create a new table or replace an existing one.
APPEND with merge key
If this mode is set, the uploaded data is merged with already persisted table data on the server. In merge mode, new data is appended to the table in additional rows and existing rows of the table will be overwritten individually only if the corresponding row of the uploaded table is more recent. At the end of the data upload, the source table contains all the data.
The ingestion API uses a merge key to merge existing data and new data in the source table. You can use the merge mode by setting the merge key in the source table definition. You also must set the persistence mode (persistenceMode) to APPEND in the source table definition.
To configure a merge key for an existing table, you must send a source table definition with the new merge key to the server. To do this, you can use the Create or replace source tables endpoint. The merge key is automatically added to a source table when the source table is created or replaced.
Note
A source table with the OVERWRITE persistence mode corresponds to a standard table in ARIS Process Mining and a source table with the APPEND consistence mode and with a specified merge key corresponds to an incremental table accordingly.