CODEWITHSUNDEEP
pythonpandaslatex
gsmafra
gsmafra

Reputation: 2504

Export pandas DataFrame to LaTeX and apply formatters by row

I want to export some DataFrames to LaTeX but these DataFrames have lots of columns, but not that many items. The solution is to display the table transposed. I know about pandas' transpose, but I want to format my rows and the method to_latex supports only formatters by column.

Some code:

import pandas as pd

names = ['Bob','Jessica','Mary','John','Mel']
births = [968, 155, 77, 578, 973]
tralala = [0.1, 0.2, 0.3, 0.4, 0.5]
BabyDataSet = zip(names,births,tralala)
df = pd.DataFrame(data = BabyDataSet, columns=['Names', 'Births', 'Whatever'])
df = df.transpose()
print df.to_latex()

Output:

\begin{tabular}{llllll}
\toprule
{} &    0 &        1 &     2 &     3 &    4 \\
\midrule
Names    &  Bob &  Jessica &  Mary &  John &  Mel \\
Births   &  968 &      155 &    77 &   578 &  973 \\
Whatever &  0.1 &      0.2 &   0.3 &   0.4 &  0.5 \\
\bottomrule
\end{tabular}

But what if I want to this for example:

\begin{tabular}{llllll}
\toprule
{} &    0 &        1 &     2 &     3 &    4 \\
\midrule
Names    &  Bob   &  Jessica &  Mary &  John &  Mel \\
Births   &  9.7e2 &    1.6e2 & 7.7e1 & 5.8e2 &  9.7e2 \\
Whatever &  0.1   &      0.2 &   0.3 &   0.4 &  0.5 \\
\bottomrule
\end{tabular}

Is there any way to hack this functionality? Only thing I thought about was to manually apply the formats converting all my columns to strings before transposing and exporting

Upvotes: 11

Views: 3749

Answers (1)

gsmafra
gsmafra

Reputation: 2504

I came up with this:

for key in df.columns.values:
    if key in formatters:
        df[key] = df[key].apply(formatters[key])
print df.to_latex()

This is pretty much equivalent to

print df.to_latex(formatters=formatters)

and works with transposed DataFrames as well i.e. df.T.to_latex()

Upvotes: 2

Related Questions