[PYTHON] Jupyter tricks

what is this

In Python3 Jupyter (IPython notebook), I will introduce the following method.

--Display pandas large DataFrame without omitting it --Deployment of generator -Wide the width of Notebook --Hide DataFrame index

Preparation

First, install more-itertools.

bash

pip install more-itertools

Then do the following: Alternatively, place a file with the following content in <user's directory> /.ipython/profile_default/startup.

jupyter

import IPython.core.getipython
from more_itertools import chunked
ip = IPython.core.getipython.get_ipython()
class ChunkedDataFrame:
    def __init__(self, *args): self.df, self.n, self.m = args
    def _repr_html_(self):
        return ('' if not isinstance(self.df, pd.DataFrame) else
            '\n'.join(self.df.iloc[:self.m, r]._repr_html_()
                      for r in chunked(range(self.df.shape[1]), self.n)))
def C_impl(self, args):
    return ChunkedDataFrame(_, *[int(i) for i in (args+' 12 5').split()[:2]])
def L_impl(self, args):
    print(list(_))
if ip:
    ip.define_magic('C', C_impl)
    ip.define_magic('L', L_impl)

Display large pandas DataFrame without omitting it

Let's display a dummy DataFrame.

jupyter

import numpy as np, pandas as pd
a = pd.DataFrame(np.random.randint(0, 100, (3, 24)))
a

The middle is omitted. Following the DataFrame display, execute as follows.

jupyter

%C

It is displayed in 12 columns at a time. If you want to change the number of columns, you can specify it as "% C 8". Let's display the DataFrame again and then try again.

jupyter

a

jupyter

%C 8

You can also change the number of lines with the second argument.

jupyter

a

jupyter

%C 8 2

Deploying the generator

If you look at the generator as it is, you cannot see the contents. It would be nice to use "list (_)", but I tried to use only 2 characters.

jupyter

(i for i in range(10))
>>>
<generator object <genexpr> at 0x000000000XXXX>

jupyter

%L
>>>
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

Widen the width of the notebook

jupyter

%%HTML
<style>
    div#notebook-container    { width: 95%; }
    div#menubar-container     { width: 65%; }
    div#maintoolbar-container { width: 99%; }
</style>

Hide DataFrame index

jupyter

df.style.set_table_styles([{'selector': 'tbody th', 'props': [('visibility','hidden')]}])

reference

―― In IPython, when I tried to see the value, it was a generator, so I came up with it when I was frustrated. -Prevent the display from being omitted in pandas

that's all