When working on #2722, I realized most PyGMT functions/methods don't allow a list of file names as input. For example,

import pygmt
fig = pygmt.Figure()
fig.plot(data=["@hotspots.txt", "@Table_5_11.txt"])
pygmt-session [WARNING]: Unable to parse 2 longitude strings

The main reason is that the data_kind function incorrectly returns a matrix type:

>>> from pygmt.helpers import data_kind
>>> data_kind(data=["@hotspots.txt", "@Table_5_11.txt"])
'matrix'

then we call the virtualfile_from_vectors to pass the data to the GMT API, which is expected to fail.

Then, the question is, if a list is passed to data, how do we know if it's a list of files or a valid record (e.g., data=[0, 1]).

To answer the question, we need to have a strict definition for "table files". As defined (based on my personal understanding) in #2722,

• A table is a 2D array of data, with M rows and N columns. Each column
  represents a different variable (e.g., x, y and z) and each row
  represents a different record.

Under this definition, a table should always be a 2D array/list, even if it has only one row. Thus, data=[[0, 1]] is a valid table, and data=[0, 1] is not).

If we agree with the strict definition, then the solution is to check if data is 1-D or 2-D array. If 1-D, then it's a list of file names, otherwise, it's a 2-D table.

With this solution, it also means the issue report #1132 is no longer valid.