BUG: DataFrame does not display as LaTeX in Jupyter when it should

[AllenDowney]

• [ X] I have checked that this issue has not already been reported.

• [ X] I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

If you run the following in a Jupyter cell:

import pandas as pd

pd.options.display.latex.repr = True

d = {'one' : [1., 2., 3., 4.],
     'two' : [4., 3., 2., 1.]}
df = pd.DataFrame(d)
df

The result is the HTML repr, as usual. But if you select a column, the Series shows the LaTeX repr, as it should.

df['one']

But DataFrame has _repr_latex_, because this works:

df._repr_latex_()

So it seems like DataFrame is ignoring the display.latex.repr option.

Problem description

When display.latex.repr is True, DataFrames should get represented in LaTeX, not HTML.

Expected Output

'\\begin{tabular}{lrr}\n\\toprule\n{} &  one &  two \\\\\n\\midrule\n0 &  1.0 &  4.0 \\\\\n1 &  2.0 &  3.0 \\\\\n2 &  3.0 &  2.0 \\\\\n3 &  4.0 &  1.0 \\\\\n\\bottomrule\n\\end{tabular}\n'

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit : 7d32926 python : 3.7.9.final.0 python-bits : 64 OS : Linux OS-release : 4.15.0-72-generic Version : #81~16.04.1-Ubuntu SMP Tue Nov 26 16:34:21 UTC 2019 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 1.2.2
numpy : 1.19.2
pytz : 2021.1
dateutil : 2.8.1
pip : 21.0.1
setuptools : 52.0.0.post20210125
Cython : None
pytest : 6.2.2
hypothesis : None
sphinx : 3.4.3
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.6.2
html5lib : 1.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.20.0
pandas_datareader: None
bs4 : 4.9.3
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.3.4
numexpr : 2.7.2
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.6.0
sqlalchemy : 1.3.23
tables : 3.6.1
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : None

[AllenDowney]

In case anyone else runs into this problem, I am currently working around it with this:

pd.DataFrame._repr_html_ = pd.DataFrame._repr_latex_

If it's stupid and it works...

[westurner]

Should Pandas instead return a mime bundle with one or more representations of the table?

• ENH: return mime bundles for Jupyter: text/plain, text/html, text/latex
  • Index?
  • Series
  • DataFrame
  • Panel
• CFG: configurable list of formats to include in mime bundles for Jupyter

From https://news.ycombinator.com/item?id=25923123 :

A display_data Jupyter kernel message includes a data key with a dict value: "The data dict contains key/value pairs, where the keys are MIME types and the values are the raw data of the representation in that format." https://jupyter-client.readthedocs.io/en/latest/messaging.html#display-data

This looks like it does something with MIME bundles: https://github.com/jupyter-xeus/xeus-cling/blob/00b1fa69d17b390b7d3b470a9859281350863d30/src/xmime_internal.hpp#L173

ipython.display: https://github.com/ipython/ipython/blob/master/IPython/display.py

ipython.core.display: https://github.com/ipython/ipython/blob/master/IPython/core/display.py

ipython.lib.display: https://github.com/ipython/ipython/blob/master/IPython/lib/display.py

[westurner]

https://ipython.readthedocs.io/en/stable/api/generated/IPython.core.formatters.html?highlight=text%2Flatex#IPython.core.formatters.DisplayFormatter :

Possible MIME types:

text/plain
text/html
text/markdown
text/latex
application/json
application/javascript
application/pdf
image/png
image/jpeg
image/svg+xml

Which of these formats does pandas produce now?

• text/plain -- __str__()
• text/html -- _repr_html_()
• text/markdown
• text/latex -- _repr_latex_()
• application/json
• application/ld+json
• application/pdf

[attack68]

I looked into MIME bundles recently as part of the PR above. One concern was performance. If Styler is processing a large DataFrame which may take a few seconds for HTML this may take longer for LaTeX and more for other types.
The implementation above yield either repr_html or repr_latex depending on the pandas option.

[attack68]

We are working toward implementing the Styler version of to_latex by default which doesn't have the problem you addressed after the above merge