From https://www.docusign.com/blog/developers/tabs-deep-dive-anatomy-tab the author writes:

“For now, it’s important to note that the ‘documentId’ and ‘pageNumber’ properties are only required if you are not placing your tabs with anchor strings.”

The above makes sense if there is only one document to sign. However if multiple documents need to be signed how does this work?

Let’s assume that I need two documents signed by someone. I construct the following information to be sent via the API:

{
    "emailSubject": "Please sign attached doc",
    "emailBlurb": "Info to be signed",
    "documents": [
        {
            "documentId": "1",
            "documentBase64": "Base64 string",
            "name": "Testfile1.pdf"
        },
        {
            "documentId": "2",
            "documentBase64": "Base64 string",
            "name": "Testfile2.pdf"
        }

    ],
    "recipients": {
        "signers": [
            {
                "email": "[email protected]",
                "name": "John Doe",
                "recipientId": "1",
                "routingOrder": "1",
                "tabs": {
                    "signHereTabs": [
                        {
                            "anchorString": "[POS_A]",
                            "documentId": "1"
                        },
                            "anchorString": "[POS_B]",
                            "documentId": "2"
                        }
                    ]
                }
            }
        ]
    },
    "status": "sent"
}

Before reading the aforementioned documentation I assumed that I could specify the ‘documentId’ property of the ‘signhereTabs’ tab, which in turn would mean that the ‘anchorstring’
property value of ‘[POS_A]’ would be searched for in the document with ‘documentId’=1 and the ‘anchorstring’ property value of ‘[POS_B]’ would be searched for in the document with ‘documentId’=2.

However, As I now understand: the two ‘documentId’ properties are irrelevant AND both documents are searched for for the two ‘anchorstring’ property values ‘[POS_A]’ and ‘[POS_B]’

Is this correct?

Furthermore, what happens if both documents contain the two tags ‘[POS_A]’ and ‘[POS_B]’?