如何强制绘图在python中显示x轴值

Question

I have an issue making a simple plot while setting the x-axis in python. Here is my code:

import import matplotlib.pyplot as plt

y  = [2586.087776040828,2285.8044466570227,1991.0556336526986,1719.7261325405243,1479.8272625661773,1272.5176077500348,1096.4367842436593,949.02201512882527,826.89866676342137,726.37921828890637,636.07392349697909,553.52559247838076,480.71257022562935,418.00424110010181,364.41801903538288,318.67575156686001,280.17668207838426,248.15399589447813,221.75070551820284,199.59983992701842,179.72014852370447,162.27141772637697,147.14507926321306,134.22828323366301,123.36572367962557,114.33589702168332,106.8825327470323,100.69181027167537,95.515144406404971,91.091036326792434]
x = range(0,30)

fig3_4 ,ax3_4 = plt.subplots()
ax3_4.semilogx(range(0,30),(loss_ave_hist))
ax3_4.set_title('Validation Performance')
# ax3_4.set_xticks(np.arange(0,30, 1.0))
ax3_4.set_xlabel('i')
ax3_4.set_ylabel('Average Loss')
fig3_4.show()
plt.show()

I believe my code is right! Notice the line of code I commented, it should set the axis with the values I want, however, it throws an error. I cannot figure out why!

Here is my plot from the my plot:

Answer 1

I used the following and it ran without errors.

All I changed is the typo in your first line of your imports, and replaced loss_ave_hist with y (ie what you called your data in your question.

y  = [2586.087776040828,2285.8044466570227,1991.0556336526986,1719.7261325405243,1479.8272625661773,1272.5176077500348,1096.4367842436593,949.02201512882527,826.89866676342137,726.37921828890637,636.07392349697909,553.52559247838076,480.71257022562935,418.00424110010181,364.41801903538288,318.67575156686001,280.17668207838426,248.15399589447813,221.75070551820284,199.59983992701842,179.72014852370447,162.27141772637697,147.14507926321306,134.22828323366301,123.36572367962557,114.33589702168332,106.8825327470323,100.69181027167537,95.515144406404971,91.091036326792434]   

import matplotlib.pyplot as plt
fig3_4 ,ax3_4 = plt.subplots()
x = range(0,30)
ax3_4.semilogx(range(0,30),(y))
ax3_4.set_title('Validation Performance')
# ax3_4.set_xticks(np.arange(0,30, 1.0))
ax3_4.set_xlabel('i')
ax3_4.set_ylabel('Average Loss')
plt.show()

UPDATE : I understand you want to label the x-axis with values from 0..29, but on a log scale, all those numbers are very close.

Here is an image with xticks set (I din't get any errors):

fig3_4 ,ax3_4 = plt.subplots()
x = range(0,30)
ax3_4.semilogx(range(0,30),(y))
ax3_4.set_title('Validation Performance')
ax3_4.set_xticks(np.arange(0,30, 1.0))
ax3_4.set_xlabel('i')
ax3_4.set_ylabel('Average Loss')
plt.show()

Here is an image where I replace semilogx with semilogy .

fig3_4 ,ax3_4 = plt.subplots()
x = range(0,30)
ax3_4.semilogy(range(0,30),(y))
ax3_4.set_title('Validation Performance')
ax3_4.set_xticks(np.arange(0,30, 1.0))
ax3_4.set_xlabel('i')
ax3_4.set_ylabel('Average Loss')
plt.show()

Does any of this resemble your goal?

Answer 2

Here is a way to make a semilogx plot but with xticks labelled according to their original (non-log) values.

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker

y = np.array([2586.087776040828, 2285.8044466570227, 1991.0556336526986, 1719.7261325405243, 1479.8272625661773, 1272.5176077500348, 1096.4367842436593, 949.02201512882527, 826.89866676342137, 726.37921828890637, 636.07392349697909, 553.52559247838076, 480.71257022562935, 418.00424110010181, 364.41801903538288, 318.67575156686001, 280.17668207838426, 248.15399589447813, 221.75070551820284, 199.59983992701842, 179.72014852370447, 162.27141772637697, 147.14507926321306, 134.22828323366301, 123.36572367962557, 114.33589702168332, 106.8825327470323, 100.69181027167537, 95.515144406404971, 91.091036326792434])
x = np.arange(1, len(y)+1)

fig, ax = plt.subplots()
ax.plot(x, y, 'o-')
ax.set_xlim(x.min(), x.max())
ax.set_xscale('log')
formatter = mticker.ScalarFormatter()
ax.xaxis.set_major_formatter(formatter)
ax.xaxis.set_major_locator(mticker.FixedLocator(np.arange(0, x.max()+1, 5)))
plt.show()

yields FixedLocator(np.arange(0, x.max()+1, 5))) places a tick mark at every 5th value in x . With ax.xaxis.set_major_locator(mticker.FixedLocator(x)) , the xticklabels got a bit too crowded.

---

Note I changed x = range(0, 30) to x = np.arange(1, len(y)+1) since the length of x should match the length of y and since we are using a logarithmic x -axis, it does not make sense to start at x=0 .

Notice also that in your original code the first y value (2586.08...) is missing since its associated x value, 0, is off-the-chart on a logarithmic scale.