Update test to reflect charset-normalizer updates

We don't pin charset-normalizer versions, so this test is likely to
flake as changes are made to that library.

- charset-normalizer 3.3.0 reports this encoding as CP949.
- charset-normalizer 3.3.1 and 3.3.1 report it as Big5.

Perhaps we don't care what the value is here. Should we remove this
test? Or, perhaps change it to assert that the encoding is not utf-8?

Author: harto

tests/test_utilities.py

@@ -54,7 +54,7 @@ def test_empty_df():
     [
         (b"abcde", "utf-8"),  # straight up ascii is a subset of unicode
         (b"Eyjafjallaj\xc3\xb6kull", "utf-8"),  # actual unicode
-        (b"\xC4pple", "cp037"),  # non-unicode, ISO characterset
+        (b"\xC4pple", "Big5"),  # non-unicode, ISO characterset
     ],
 )
 def test_detect_encoding(test_string, encoding):