迭代绘制堆积直方图pandas / matplotlib

Question

I'd like to iterate through the columns of a dataframe and plot a stacked histogram for each column that distinguishes between two groups (where death = 0 vs 1). How do I convert this code to something iterative? (bun_max is one of the columns used as an example.) (Also, how do I get the legend to work?)

df1 = temp[temp['death'] == 0]
df2 = temp[temp['death'] == 1]

plt.figure()
plt.hist([df1.bun_max, df2.bun_max], bins=50, stacked=True, color=['b','r']);
plt.title(df1.bun_max.name)
plt.ylabel('ICU admits')
plt.xlabel(df1.bun_max.name)
plt.legend()
plt.show()

Example output

This is what I have so far. I get an error: "TypeError: len() of unsized object". All columns are either int or float. Would be helpful to understand the reason for the error.

for x in df1:
    for y in df2:
        plt.figure()
        plt.hist([x, y], bins=50, stacked=True, color=['b','r'])
        plt.title(df1.x.name)
        plt.show()

TypeError: len() of unsized object

Answer 1

I figured it out:

df1 = temp[temp['death'] == 0]
df2 = temp[temp['death'] == 1]
df1 = df1.drop('death', axis=1)
df2 = df2.drop('death', axis=1)

for col1 in df1.columns:
    for col2 in df2.columns:
        if col1 == col2:
            plt.figure();
            plt.hist([df1[col1], df2[col2]], bins=50, stacked=True, color=['b','r']);
            plt.title(df1[col1].name)
            plt.ylabel('ICU admits')
            plt.xlabel(df1[col1].name)
            plt.show();