我正在尝试使用pandas（csv）的to_csv()函数将以下数据帧转换为v0.24.0文件：
robustness orseilline disrupt
jarkman woodprint antennariid
sublinguae 436 47588 60
coupee 994 67199 21
Sauropsida 494 97139 57
metapleural 652 85448 95
sulphuret 667 55142 96
行的顺序很重要，但是结果会自动按字母顺序对行进行排序：
,robustness,orseilline,disrupt
,jarkman,woodprint,antennariid
Sauropsida,436,47588,60
coupee,994,67199,21
metapleural,494,97139,57
sublinguae,652,85448,95
sulphuret,667,55142,96
DataFrame是使用一些可创建随机单元格的代码生成的：
# 'rows' is predefined by some other generating function
stubs = [random.sample(words.words(), self.n_rows) for x in range(self.n_stubs)]
stubs = pd.MultiIndex.from_arrays(stubs)
headers = [random.sample(words.words(), self.n_columns) for x in range(self.n_headers)]
headers = pd.MultiIndex.from_arrays(headers)
table = pd.DataFrame(data=rows, index=stubs, columns=headers)
当将DataFrame转换为csv时，MultiIndex显然具有一些顺序（称为“代码”）。使用.to_latex()或to_html()时，顺序被保留。该如何处理？