DataFrame.to_latex() should honor na_rep after formatter.

[moorepants]

In the example below I would expect that 'np.nan' would be replaced by 'aaah'.

In [3]: df = pd.DataFrame({'a': np.arange(3), 'b': [0.1, np.nan, 0.2]})

In [4]: df
Out[4]: 
   a    b
0  0  0.1
1  1  NaN
2  2  0.2

In [5]: print(df.to_latex(formatters={'b': lambda x: '{:0.4f}'.format(x)}, na_rep='aaah'))
\begin{tabular}{lrr}
\toprule
{} &  a &      b \\
\midrule
0 &  0 & 0.1000 \\
1 &  1 &    nan \\
2 &  2 & 0.2000 \\
\bottomrule
\end{tabular}

[jreback]

this is not implemented AFAICT. pull-requests are welcomed.