Multilevel indexing with .to_latex() method merges two index columns

[jobingr]

using Release 0.10.1
using the .to_latex method on a multilevel frame (specifically using a pivot_table) merges the two index columns so that it does not render correctly in latex. Looks like it is missing the & for blank cells in index. and also an & for the columns. Manually inserted those in output and it fixes problem

[jobingr]

PS sorry definitely wrong place for this, but any idea on where to find documentation on .to_latex method. Only info I can find is in docstring and can't figure out how to use it to format numbers and alignments. considering regexes, but I have feeling the functionality is already there, just cant figure it out.

[ghost]

You might be looking for formatters:

http://pandas.pydata.org/pandas-docs/dev/io.html?highlight=formatters#writing-a-formatted-string

[jobingr]

had look at the formatters can' t figure it out Tried:
form={'colname': round()} as index also tried the formula in quotes. I am missing something? is it a string formatter for the numbers?

[ghost]

round() calls round and puts the result as the formatter, not what you want.
try round (no parens).

[jobingr]

thanks.. any suggestions for currency formatting... no need for symbol, but thousand separators will make it much more readable. Also is there a way to change the default c of the to_latex method to r. can manualy munge the latex output, but I suspect there is already a builtin way of doing this.

[ghost]

I think python .format() style string formatting does thousands.
The 'c' is hard-coded I believe, so no to the latter. Haven't done latex
in a long time so I'm not sure what 'c' vs. 'r' actually does, but if you need
it maybe others do too, Pull Requests are welcome.

[jobingr]

tried:
x=randn(10)
def f(x):
return round(3)
fr = pd.DataFrame(x, columns=['test'])
fr.to_latex(formatters={'test': f})
also tried f with () still error

[jobingr]

c makes column centre justified, can use l or r to to left and right. numbers more readable if r justified. that is an easy fix though.

[jobingr]

got it figured formatter needs string return
import pandas as pd
x=randn(10)
def f(x):
return str(round(x,3))
fr = pd.DataFrame(x, columns=['test'])
fr.to_latex(formatters={'test': f})
worked
thanks for help

[jobingr]

Got split up here for a while... will close this one an re-add the issue with to_latex method on multi indexed frames on its own