5

I have a (very simplified) segment of code here for an open source project I'm working on for plotting ATE results from STDF data...

def plot_everything_from_one_test(test_data, data, num_of_sites, test_tuple):

    plt.figure()


    # Title for everything
    plt.suptitle("Test Information Goes Here")

    # Find the limits
    low_lim = get_plot_min(data, test_tuple, num_of_sites)
    hi_lim = get_plot_max(data, test_tuple, num_of_sites)

    ######        

    table = function_that_returns_a_dataframe()

    plt.subplot(121)

    # HOW WOULD I ACTUALLY PUT A DATAFRAME HERE

    ######

    # Plots the trendline
    plt.subplot(222)
    plot_full_test_trend(test_data, low_lim, hi_lim)
    plt.title("Trendline")

    # Plots the histogram
    plt.subplot(224)
    plot_full_test_hist(test_data, low_lim, hi_lim)
    plt.title("Histogram")

    plt.show()

Basically, all I really want to do is in subplot section (121) to render simply a table containing the contents of an array or dataframe or whatever. This dataframe happens to be 10 columns wide and indefinitely long (in the order of magnitude of ~10s, but can be trimmed if needed). I've looked around and I know matplotlib has table(), but from what I can see it creates a new figure (unwanted) and it also tries to force a graph to render on top of it, even though I really don't want to have that render because it really would be useless. I saw this could be sort-of circumvented, but I don't know how I would manage it while preserving my subplots. Perhaps I just don't understand how matplotlib works with them.

Another problem is that when I can sort-of get tables to render, they are absurdly SLOW, reaching upwards of a minute to render a 10x30-something dataframe. Also, the text renders absurdly too small and doesn't scale. Sorry if this all sounds like a cluster of problems, but I really don't know how I would deal with what should be a simple render of a table. I may just convert to a string if I really need and render it as text. Thanks in advance!

McNibbler
  • 59
  • 1
  • 1
  • 2
  • `table` does not create a new figue. It also does not force anything on top of it. There are many questions on tables around. So it would be good if you share the code you have a problem with and tell explicitely why those other questions do not help, see [ask] and [mcve]. Same for the timing - without seeing what you have done in terms of code, one can only speculate on why it takes so long. – ImportanceOfBeingErnest Mar 09 '18 at 20:03

1 Answers1

9

You can add a table to a subplot - you dont need a separate figure for it. Then you can turn the axis off for that subplot and it wont try to render anything other than the table in that region

You also mentioned that your code is running slow when trying to render a 10x30 dataframe in the subplot. There might be an issue with you implementation of the table because the following renders within a second:

import matplotlib.pyplot as plt
import pandas as pd

d = {'x{}'.format(i): range(30) for i in range(10)}

table = pd.DataFrame(d)

plt.figure()

# table
plt.subplot(121)

cell_text = []
for row in range(len(table)):
    cell_text.append(table.iloc[row])

plt.table(cellText=cell_text, colLabels=table.columns, loc='center')
plt.axis('off')

# plot
plt.subplot(222)
plt.plot(table['x1'],table['x2'])

# plot
plt.subplot(224)
plt.plot(table['x1'],table['x2'])

plt.show()

enter image description here

Simon
  • 9,762
  • 15
  • 62
  • 119