How do you suppress matplotlib output in for loop in Jupyter Notebook?

Question

I am looping over a list of DataFrame column names to create barplots using matplotlib.pyplot in a Jupyter Notebook. Each iteration, I'm using the columns to group the bars. Like so:

%matplotlib inline

import pandas as pd
from matplotlib import pyplot as plt


# Run all output interactively
from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"


df = pd.DataFrame({'col1': ['A', 'B', 'C'], 'col2': ['X', 'Y', 'Z'], 'col3': [10, 20, 30]})

#This DOES NOT suppress output
cols_to_plot = ['col1', 'col2']
for col in cols_to_plot:
    fig, ax = plt.subplots()
    ax.bar(df[col], df['col3'])
    plt.show();

The semicolon (';') should suppress text output but when I run the code I get, after the 1st run:

If I run a similar snippet outside of a for loop, it works as expected - the following successfully suppresses output:

# This DOES suppress output
fig, ax = plt.subplots()
ax.bar(df['col1'], df['col3'])
plt.show();

How do I suppress this text output when looping?

Note:

In a previous version of this question, I used the following code to which some of the comments refer, yet I changed it to the above to better show the issue.

cols_to_boxplot = ['country', 'province']
for col in cols_to_boxplot:
    fig, ax = plt.subplots(figsize = (15, 10))
    sns.boxplot(y=wine['log_price'], x=wine[col])
    labels = ax.get_xticklabels()
    ax.set_xticklabels(labels, rotation=90);
    ax.set_title('log_price vs {0}'.format(col))
    plt.show();

I cannot reproduce this. When running the above, there is no text output. But I guess you should use sns.boxplot( ..., ax=ax) to plot to the axes you create in the line above. — ImportanceOfBeingErnest
– ImportanceOfBeingErnest, Commented Mar 11, 2018 at 10:29
@ImportanceOfBeingErnest, thanks for suggestion, but that didn't work. I wonder why you can't reproduce it. Maybe it's version-specific. — voyager
– voyager, Commented Mar 11, 2018 at 10:53
My suggestion has nothing to do with this issue, it's just in general that you should tell boxplot which axes to use if you created that axes beforehands - in this case it will not break anything if you don't do it, in other cases it would. Concerning the issue: Yes, post a minimal reproducible example, tell the versions you are using and best include a screenshot as well. For me, it looks like this. — ImportanceOfBeingErnest
– ImportanceOfBeingErnest, Commented Mar 11, 2018 at 11:01
I can also not reproduce this. I just get plots in the output in the same way as the screenshot in @ImportanceOfBeingErnest's comment — jarmokivekas
– jarmokivekas, Commented Mar 11, 2018 at 12:15
Thanks for the responses. I edited my post as suggested but in doing so also answered my own question, which I also posted. — voyager
– voyager, Commented Mar 11, 2018 at 13:36

voyager · Accepted Answer · 2018-03-11 13:23:46Z

7

I discovered what what causing this behaviour. I ran my notebook with the following:

from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"

(documented here)

This had the effect of not suppressing matplotlib output when plotting in a loop. However, as mentioned in the original post, it did suppress output as expected when not within a loop. In any case, I fixed the issue by 'undoing' my code above like this:

InteractiveShell.ast_node_interactivity = "last_expr"

I'm not sure why this would happen.

answered Mar 11, 2018 at 13:23

voyager

3433 silver badges10 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Mathews24 Over a year ago

Setting it to "none" helped resolve my error

Arend · Accepted Answer · 2022-01-31 08:52:01Z

0

@voyager's answer worked for me, but only if I put

InteractiveShell.ast_node_interactivity = "last_expr"

in the cell preceding the cell with the matplotlib annotate loop.

Alternative: I wrapped the plot code containing the annotate loop into a plot function. If you call this function, the repeated 'Text' output remains within the function, no longer requiring a change to the InteractiveShell settings:

def plot_with_loop(df):
   ...
   ax = df.plot();

   for i in range(len(df)):
        # next line generates 'Text' output when not called contained in function
        ax.annotate(...); 
    

plot_with_loop(df)  # no 'Text' output

answered Jan 31, 2022 at 8:52

Arend

3332 silver badges5 bronze badges

Comments

Dominic van der Pas · Accepted Answer · 2023-11-24 12:07:33Z

0

Try replacing plt.show() with plt.close().

It worked for me.

answered Nov 24, 2023 at 12:07

Dominic van der Pas

1428 bronze badges

Collectives™ on Stack Overflow

How do you suppress matplotlib output in for loop in Jupyter Notebook?

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related