SharePoint Custom Fields (Single line of text) - Gets cast to Python Pendulam DateTime

[nikhilmakan02]

Hi Everyone, doing a simple call to a SharePoint Document library to return the custom metadata fields related to a file. What's interesting is I can set the column type to 'Single line of text' in SharePoint but if I put certain numbers into the column it returns as a datetime object.

If I put text in the field on SharePoint it returns correctly as text or even if I put enough numbers it returns as text as well.

In the example below the Doc_ID column in SharePoint contains the numbers 1212.

result = await (
        client.drives.by_drive_id(drive_id)
        .items.by_drive_item_id(drive_item_id)
        .list_item.get()
    )

print(result.fields.additional_data["Doc_x0020_ID"])
print(type(result.fields.additional_data["Doc_x0020_ID"]))

Output:

1212-01-01 00:00:00+00:00
<class 'pendulum.datetime.DateTime'>

This seems like unexpected behavior?

[georgethomasii]

I am seeing a similar issues when querying entraid attributes. Any value that has a 4-digit numerical value is being returned as a datetime object instead of a string. If I get the raw http response the value is returned as a string correctly. It would be great if this was looked into and resolved as I prefer not to have to get the native HTTPX response.

[shemogumbe]

Hello, and thanks for continued use of the SDK and for raising this.
As we look into our serialization to give a consistent behaviour, would you cast the result to a string/datetime object for now:

• Pendulum to string

# Cast Pendulum DateTime to string
pendulum_datetime_string = pendulum_datetime.to_iso8601_string()
print("Pendulum DateTime as string (ISO 8601 format):", pendulum_datetime_string)

• string to pendulum

# Parse the string to Pendulum DateTime object
pendulum_datetime = pendulum.parse(pendulum_datetime_if_string)

[nikhilmakan02]

@shemogumbe I am not really sure how that is helpful? Are you suggesting I cast this to a string then truncate the string to get the 4 digit value. That sounds like a messy solution to be honest.

I have actually just implemented @georgethomasii suggestion to just get the raw response and get the metadata more reliably there.