Template for variables in to_latex()

[eXcuvator]

Me and several of my colleagues typically create output for variables that have greek letters. Inside pandas, we then call them "epsilon", "sBar" or similar. Right now we have to run a string replace on the output of to_latex() and then manually save it to a file.

It'd be nice to be able to give to_latex() a template such as ['epsilon':'$\epsilon$', 'sBar':'$\bar s$'] and then to_latex() would replace these in the output accordingly.

[onesandzeroes]

The kind of replacements you're describing are already possible using DataFrame.map() with a dictionary:

import pandas as pd
df = pd.DataFrame({'latex_raw': ['sBar', 'epsilon']})
replacements = {'sBar': r'$\bar{s}$', 'epsilon': r'$\epsilon$'}
df['latex_formatted'] = df['latex_raw'].map(replacements)
df
# Output:
  latex_raw latex_formatted
0      sBar       $\bar{s}$
1   epsilon      $\epsilon$

So you should be able to create replaced versions of your columns before doing to_latex(). Is there any particular advantage to duplicating the map() functionality within to_latex(), any examples you can provide of things that aren't easy to do with that approach?

[TomAugspurger]

Or we can lump this in with #3190 when somebody finally gets around to it (I'll possibly take another crack for the next release).

[eXcuvator]

Since I was mainly referring to label rather than values, I suppose the alternative approach suggested is valid.