Instruction for converting csv-files, produced here by NumPy, to a format required in TeX-tables.

We will start at the point, where all data is calculated.

The line in python should be according to

numpy.savetxt('output.csv',
    numpy.column_stack([data1, data2, ...]),
    fmt='%.2f, %.0f, ...',
    header='data1, data2, ...')

The option fmt let you decide how many decimal places should be written. In this example 2 for data1 and 0 for data2.

IMPORTANT:

• Watch that you set a format for each column!
• Use the , between each of the numbers (as shown)

If you try to change the format simply from .csv to .tex it would not work. I tried to get things done with the end-of-line character option in numpy-savetxt but it did not satisfy me.

Instead we can use the two commands

sed '/^#/ d' < output.csv > output-new.csv

cat output-new.csv | sed \
    -e 's:,:\t\&\t:g' \
    -e 's:$:\t\\\\:g' > output.tex

With those command we can prepare the output.csv-file in a way, that we only need to import it into LaTeX.

The first sed command removes the header-line and writes it into a new file. (For some reason it does not work for me with the same file; if you do not use a header, you can remove the line)

Quick rundown of the arguments from second sed:

• Replacing the , as a delimiter with the & used by TeX (one could use the & already in the fmt-option)
• Replacing the end of line, withe the double backslash \\

The next part is bigger and needs to be adapted for each table, but i did not fin a way without it.

ed output.tex << END
1i
\begin{tabular}{c c ...}
\toprule
{Data 1} & {Data 2} & ... \\\\
\midrule
.
\$a
\bottomrule
\end{tabular}
.
wq
END

These lines put the previous prepared data into the tabular structure. For each table, change

• the number of columns and the way to display them (c or S)
• the title of each column

\begin{table}
  \input{output.tex}
\end{table}

Of course you should add things as centering, caption and label as needed.

You may also take a look into the example.