fix: fix rows returned when both start_index and page_size are provided

[Linchin]

TODO:

• move deep copy import
• double check total rows - done, total rows should not be affected by start_index
• unit and system tests

Fixes #2173 🦕

[chalmerlowe]

Please go through the rest of this test and provide some additional comments.

Clarify why we are using total_rows of 7.
Clarify why setting total_rows to 7 despite only have six rows of tabledata is ok/reasonable
Clarify what is at the indexed positions:

• tabledata_list_request_1[1]
• tabledata_list_request_2[1]

[Linchin]

Clarify why we are using total_rows of 7.
Clarify why setting total_rows to 7 despite only have six rows of tabledata is ok/reasonable

Done.

Clarify what is at the indexed positions:
tabledata_list_request_1[1]
tabledata_list_request_2[1]

There are comments above the code block explaining that these are the first and the second requests that the client sends to the server.

[chalmerlowe]

I am gonna drop this comment.

[chalmerlowe]

Here we mention sending page_token, but nowhere in the test do we confirm that page_token is set or that it has been sent, etc.

In fact, while it is correct, with this language in place, it seems weird that in tabledata_resource_2, pagetoken is set to None. The above language implies that we might send something besides None.

Should we revise the language slightly?
Do we mean some other argument/attribute besides page_token?

[Linchin]

In the last message, the page token will be empty because there is no following page.

[Linchin]

Also, tabledata_resource_2 is what the server returns, not what we pass to the server. The process of sending the page token happens within result = job.result(page_size=page_size, start_index=start_index) and I don't think it's realistic to expose that in unit test.

[chalmerlowe]

I am gonna drop this comment.

[chalmerlowe]

Added some comments about the docstrings and comments in the test.
The test feels complicated especially upon first read so it might benefit from a bit more commentary and potentially some rewording.