Difference¶

https://stackoverflow.com/questions/19917545/comparing-two-pandas-dataframes-for-differences

df.equals¶

Will check both dtypes and data. Return a bool. not very reliable regarding floats

df.compare¶

Will only check data. Return a dataframe. same shape and type. not reliable either (we need to consider rounding errors)

df.compare(dp).empty #return True if same

# workaround - define a round function
def round_df(df, decimal=6):
    tmp = df.select_dtypes(include=[np.number])
    df.loc[:, tmp.columns] = np.round(tmp)
    return df

if all are numbers, using np.allclose

np.allclose(df1, df2, rtol=1e-6, atol=1e-6)