Duplicate index fools type inference

[jebob]

Minimum example

import gdxpds
import pandas as pd


df1 = pd.DataFrame(
    index=[0, 1, 2],
    data={
        "A": [0, 1, 2],
        "value": [3, 4, 5],
    }
)

df2 = pd.DataFrame(
    index=[0, 0, 0],
    data={
        "A": [0, 1, 2],
        "value": [3, 4, 5],
    }
)

gdxpds.to_gdx({"df1": df1, "df2": df2}, "out.gdx")

Results in

But since gdxpds doesn't use the index, I would expect them to be the same.