If I use the

incremental append

and my cursor is

updated_at

. And I get a new row with

PK

id

in a postgres source, it would still append the row with the new id? And if I do incremental append with let’s say

id

PK

will I then loose track of updated values? Since the

id could be the same

Assuming this is not

CDC

.

@[DEPRECATED] Marcos Marx turned this message into Zendesk ticket 2509 to ensure timely resolution!

Hey @Simon Thelin, I may be misunderstanding your question but yes, for the first example: I think it would still append the row with the new id as updated records will get appended (assuming the updated_at field is different) Can you explain what you mean by lose track of updated values?

So imagine I use

id

as the incremental append value. That means, that if the row of

id

updates a value at lest say column

name

. I would then miss this update, due to the fact

id

is the same? Hence why you would use

update

. So for tables, which does not have any

update

or similar, you can’t use incremental appends. To solve that you would go with CDC. Is this correct @Saj Dider (Airbyte)?

Yes you are correct, if you use id as the cursor for incremental append and would miss updates, you will have to use another incremental method like CDC. There are also some illustrative examples in our docs: https://docs.airbyte.com/understanding-airbyte/connections/incremental-append/ Hope this helps!