Is there a way to prevent the auto sorting when using the dataframe.boxplot() function?

Question

I need to label the columns and have them displayed in the original sequence.

Here is what I did :

df['Column'] = df['Column'].map({1: 'L5', 2:'L4', 3: 'L3', 4:'L2', 5:'L1', 6:'M12', 7:'M11', 8:'M10', 9:'M9', 10:'M8', 11:'M7', 12:'M6', 13:'M5', 14:'M4'})

sorted()

df.boxplot(column=['Time'], by=['Column'])

However, the plot appears to be in no particular order

Answer 1

The plot appears in lexical order of the new column value. If you want the original order, you can force it with pd.Categorical:

d = {1: 'L5', 2:'L4', 3: 'L3', 4:'L2', 5:'L1', 6:'M12', 7:'M11', 8:'M10', 9:'M9', 10:'M8', 11:'M7', 12:'M6', 13:'M5', 14:'M4'}

df['Column'] = pd.Categorical(df['Column'].map(d), ordered=True, categories=list(d.values()))

df.boxplot(column=['Time'], by=['Column'])

Or you can just relabel the ticks:

# no map
# df['Column'] = df['Column'].map(d)

# grab the axis returned by .boxplot
ax = df.boxplot(column=['Time'], by=['Column'])

# rename the tick labels, `d` as above
ax.set_xticklabels(list(d.values()))

Either of them gives you: