Skip to main content

​
Tables and Columns

​
UPSERT Query Behavior

​
UPSERT Slider Set to ON

​
UPSERT By Id

If the key column is present in the input XML, a query is issued to the database to determine if any records exist with the matching key:
A match results in an update; no match results in an insert. If the key column is not present in the input XML, the record is always inserted:

​
UPSERT By Non-key Column

A query is issued to the database to determine if any records exist where the selected UPSERT By column contains the matching value, by selecting the primary key of the table:
The SELECT query always targets the key column(s) of the table in the SELECT statement. If there is a successful match, the data in the input XML is updated into the data source. If there is no match, the record is inserted instead.
If the selected UPSERT By column is not specified in the input, the lookup is used with a NULL value, as shown in the following query:

​
Using a Custom Query

You can also write a custom query to select the UPSERT key. This option allows for the lookup to be performed using more complex logic than just a single column match. Consider the following query:
In this example, you must select the key column(s) of the table, to provide adequate criteria for the UPDATE. The @OrderID syntax is used to reference a value from the matching input XML. You can also use static values: the quotes around 'Active' denote that this clause uses a static value. Again, a successful match of the key column is used to form an update request. If there is no match, the record is inserted.
When in doubt about the behavior of a database insertion, review the .log file for the UPSERT attempt: a complete breakdown of the logic used to guide the behavior is sent to the log file.

​
UPSERT Slider Set to OFF

When the slider is set to OFF, the connector simply inserts all queries into the database. Examining the code view of the input mapping shows that the action attribute of the table is insert:
All queries issued to the table are treated as inserts, whether or not the primary key is present. To have all received records updated against the table, you can set this action keyword to update, as shown in the following snippet:
When the action is defined as update, the application expects the key column of the table (for example, Id) to be present. If the key column is not provided, the effects of the UPDATE command can vary by target database type, so you might see more rows changed than expected, or that none are changed at all.

​
Example Arc Flow