当键不统一时，如何在不使用Pandas的情况下从嵌套字典绘制条形图? [英] How to plot a bar graph from nested dictionary without using Pandas when keys are not uniform?

问题描述

我有一个像这样的嵌套字典:

I have a nested dictionary like this:

{'uvvm': {'0-250ms': 96, '500-750ms': 2}, 'usvv': {'0-250ms': 1},
 'unsharedChunksVirtualVolume': {'0-250ms': 21}, 'ubvv': {'0-250ms': 60, '250-500ms': 2, '500-750ms': 3},
 'unbvvh': {'0-250ms': 1}, 'ssvv': {'0-250ms': 101, '250-500ms': 1},
 'sscc': {'0-250ms': 2}, 'sc': {'500-750ms': 2},
 'qvv': {'0-250ms': 200, '500-750ms': 5}, 'qas': {'0-250ms': 4}}

内部键 0-250 ， 250-500 ， 500-750 等对于所有外部键都不相同.我需要绘制不使用熊猫的条形图，因为我们需要将其作为支持包的一部分进行收集，并且不能在客户计算机中安装 pandas .请提供一种方法.

The inner keys 0-250, 250-500, 500-750 etc., are not uniform for all outer keys. I need to plot a bar graph without using pandas as we need to collect as part of support bundle and cant have pandas installed in customer machines. Please provide a way.

op = dict(list(dictLatency.items())[i:i+10])
print(op)
inner_keys = ['0-250ms', '250-500ms', '500-750ms', '750-1000ms', '1000-1250ms',
              '1250-1500ms', '1500-1750ms', '1750-2000ms', '2000-2250ms']
newDict = a:[[i, b.get(i, 0)] for i in inner_keys] for a, b in op.items()
print(newDict)

labels = op.keys()

此处 newDict 是

{'updateVirtualVolumeMetaData': [['0-250ms', 96], ['250-500ms', 0], ['500-750ms', 2], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'updateStorageProfileForVirtualVolume': [['0-250ms', 1], ['250-500ms', 0], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'unsharedChunksVirtualVolume': [['0-250ms', 21], ['250-500ms', 0], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'unbindVirtualVolume': [['0-250ms', 60], ['250-500ms', 2], ['500-750ms', 3], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'unbindAllVirtualVolumesFromHost': [['0-250ms', 1], ['250-500ms', 0], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'spaceStatsForVirtualVolume': [['0-250ms', 101], ['250-500ms', 1], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'setStorageContainerContext': [['0-250ms', 2], ['250-500ms', 0], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'setContext': [['0-250ms', 0], ['250-500ms', 0], ['500-750ms', 2], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'queryVirtualVolume': [['0-250ms', 200], ['250-500ms', 0], ['500-750ms', 5], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]], 'queryArrays': [['0-250ms', 4], ['250-500ms', 0], ['500-750ms', 0], ['750-1000ms', 0], ['1000-1250ms', 0], ['1250-1500ms', 0], ['1500-1750ms', 0], ['1750-2000ms', 0], ['2000-2250ms', 0]]}

它在列表 [] 中显示值，而不是字典 {} 中的值.我希望将内键和值作为dict而不是列表.

It is showing values in list [] and not as dictionary {}. I'd like the inner key and values as dict not as list.

推荐答案

您可以遍历外部词典，然后遍历内部词典绘制堆叠的条形图:

You could loop through the outer dictionary, and then through the inner one to draw a stacked bar graph:

import matplotlib.pyplot as plt
from matplotlib.patches import Rectangle
import numpy as np

dictLatency = {'uvvm': {'0-250ms': 96, '500-750ms': 2}, 'usvv': {'0-250ms': 1},
               'unsharedChunksVirtualVolume': {'0-250ms': 21}, 'ubvv': {'0-250ms': 60, '250-500ms': 2, '500-750ms': 3},
               'unbvvh': {'0-250ms': 1}, 'ssvv': {'0-250ms': 101, '250-500ms': 1},
               'sscc': {'0-250ms': 2}, 'sc': {'500-750ms': 2},
               'qvv': {'0-250ms': 200, '500-750ms': 5}, 'qas': {'0-250ms': 4}}
inner_keys = ['0-250ms', '250-500ms', '500-750ms', '750-1000ms', '1000-1250ms',
              '1250-1500ms', '1500-1750ms', '1750-2000ms', '2000-2250ms']
# colors = plt.cm.Set3.colors
colors = plt.cm.hsv(np.linspace(0, 0.8, len(inner_keys)))

fig, ax = plt.subplots(figsize=(15, 4))
for outerkey, innerdict in dictLatency.items():
    bottom = 0
    if outerkey == 'unsharedChunksVirtualVolume':
        outerkey = 'unshared\nChunks\nVirtual\nVolume'
    for innerkey, value in innerdict.items():
        ax.bar([outerkey], [value], bottom=bottom, color=colors[inner_keys.index(innerkey)])
        bottom += value
ax.tick_params(axis='x', rotation=90)
legend_handles = [Rectangle((0, 0), 0, 0, color=color, label=label) for color, label in zip(colors, inner_keys)]
ax.legend(handles=legend_handles)
plt.tight_layout()
plt.show()

PS:对于单杠，您可以将对 ax.bar 的调用更改为:

PS: For horizontal bars you could change the call to ax.bar to:

ax.barh(y=outerkey, width=value, left=bottom, color=colors[inner_keys.index(innerkey)])

这篇关于当键不统一时，如何在不使用Pandas的情况下从嵌套字典绘制条形图?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！