[BUG] Triangle-to-Triangle comparison fails

[genedan]

Are you on the latest chainladder version?

• Yes, this bug occurs on the latest version.

Describe the bug in words

A comparison such as raa < raa fails.

How can the bug be reproduced?

import chainladder as cl
raa = cl.load_sample('raa')

raa < raa
Traceback (most recent call last):
 ... 
line 524, in __lt__
    obj.values = xp.nan_to_num(obj.values) < xp.nan_to_num(value)
                                             ~~~~~~~~~~~~~^^^^^^^
TypeError: no implementation found for 'numpy.nan_to_num' on types that implement __array_function__: [<class 'chainladder.core.triangle.Triangle'>]

What is the expected behavior?

This should behave like the numpy and Pandas analogues:

import numpy as np

a = np.array([[1.0, 2.0], [3.0, np.nan]])
b = np.array([[1.0, 1.0], [9.0, 9.0]])

print(a)
array([[ 1.,  2.],
       [ 3., nan]])

print(b)

array([[1., 1.],
       [9., 9.]])

a < b
# array([[False, False],
#        [ True, False]])      NaN compares False

a <= b
# array([[ True, False],
#        [ True, False]])      differs from < at the equal cell

a < np.array([[1.0, 2.0, 3.0]])
# ValueError: operands could not be broadcast together with shapes (2,2) (1,3)


import numpy as np
import pandas as pd

d1 = pd.DataFrame({"x": [1.0, 3.0], "y": [2.0, np.nan]}, index=["a", "b"])
d2 = pd.DataFrame({"x": [1.0, 9.0], "y": [1.0, 9.0]}, index=["a", "b"])

print(d1)
     x    y
a  1.0  2.0
b  3.0  NaN

print(d2)
     x    y
a  1.0  1.0
b  9.0  9.0

d1 < d2
#        x      y
# a  False  False
# b   True  False       NaN compares False

d1 <= d2
#       x      y
# a  True  False        differs from < at the equal cell
# b  True  False

d3 = pd.DataFrame({"x": [1.0, 9.0]}, index=["a", "b"])

d1 < d3
#       x   y
# a   2.0 NaN
# b  12.0 NaN

Would you be willing to contribute this ticket?

• Yes, absolutely!
• Yes, but I would like some help.
• No.

[henrydingliu]

@genedan I believe you are giving conflicting expected behavior between this issue and #1392. In this issue, NaN should always compare to false. In #1392, NaN is treated as 0, and can compare to True. Please let me know if I'm understanding these issues as you intend

@Abhayindia FYI