The documentation for grdtrack states that an argument for newcolname is required if the argument for points is a DataFrame. However, a GMTInvalidInput error requiring an argument for newcolname is raised if points is a numpy array or a list as well.

Full code that generated the error
Import statement

import pygmt
import numpy as np
grid = pygmt.datasets.load_earth_relief(region=[-5, 5, -5, 5])

Points as a numpy array

points_matrix = np.array([[1,1], [1,2], [2,2]])
assert type(points_matrix) == np.ndarray
grdtrack_output = pygmt.grdtrack(points=points_matrix, grid=grid)

Points as a list

points_list = [[1,1], [1,2], [2,2]]
assert type(points_list) == list
grdtrack_output = pygmt.grdtrack(points=points_list, grid=grid)

Full error message

GMTInvalidInput: Please pass in a str to 'newcolname'

The error is raised in lines 251-252 in grdtrack.py

if data_kind(points) == "matrix" and newcolname is None:
    raise GMTInvalidInput("Please pass in a str to 'newcolname'")

It looks like data_kind() won't distinguish between a DataFrame and a numpy array and returns "matrix" in both instances. Should newcolname be a required parameter for all matrix-type data, since the function still returns a DataFrame if no outfile is specified, or should the error handling check to make sure that points is a DataFrame?