Pandasでカラム名がTupleになってしまうというバグ？を踏み抜いた件

import pandas as pd

empty_df = pd.DataFrame()
print("Plain1:", empty_df)
"""
ただの空のDataFrame
Plain1: Empty DataFrame
Columns: []
Index: []
"""
empty_df["new_column"] = 1
print("Plain2:",empty_df)
"""
# Printed
Plain2: Empty DataFrame
Columns: [new_column] # <- ちゃんとstring
Index: []
"""

import pandas as pd

empty_df = pd.DataFrame(columns=["id","key"])
empty_df["value"] = 1 # ただのemptyなdfだと、pivotする時に例外が発生。これをすると何故かpivotが可能に
print("Pivot1:", empty_df)
"""
Pivot1: Empty DataFrame
Columns: [id, key, value]
Index: []
"""

pivot = pd.pivot_table(empty_df, index="id", columns="key", values="value")
print("Pivot1:", pivot)
"""
# ただの空のDataFrame
Pivot2: Empty DataFrame
Columns: []
Index: []
"""

pivot["new_column"] = 1
print("Pivot3:",pivot)
"""
# Printed
Pivot3: Empty DataFrame
Columns: [(new_column, )] # <- カラム名がなぜかTuple
Index: []
"""