应用size（）之后如何在DataFrame Groupby上进行迭代？

Question

Combing thru log files I build a dataframe of the process that failed, the date and the machine. My goal is to provide a bar chart for each process where the dates are the x-axis and the count of failures each day is computed with .size().

grouped = fail_df.groupby(['Process', 'Date']).size
print(fail_df.groupby(['Process', 'Date']).size())

shows exactly what I want. First lines of the print are

Process          Date
10HzTail         2019-06-16     1
1553Prox         2019-06-16     3
                 2019-06-17     8
                 2019-06-18    10
                 2019-06-19     2
                 2019-06-20     5
Cthread2         2019-06-18     1
                 2019-06-20     1

I try to iterate as

for name, row in grouped:
    print(name)
    print(row)

Gives this error output

dtype: int64
Traceback (most recent call last):
  File "./allpandas", line 140, in <module>
main()
 File "./allpandas", line 125, in main
  for name, row in grouped:
TypeError: 'int' object is not iterable

I would want to process each Process in turn. I want the dates and count fed to the bar chart for that Process.

Is there a way to iterate over this or have I made a fundamental mistake in my grouping?

UPDATE

I tried the suggested size() and still get the same error.

grouped = fail_df.groupby(['Process', 'Date']).size()
for name, row in grouped:
    print(name)
    print(row)

Are there other suggestions?

Answer 1

Fastest solution if you are using matplotlib

Are you using matplotlib ?
If so, if I understood what you want, you don't need to loop, you can use pandas.DataFrame.plot which does all the job for you.

grouped = fail_df.groupby(['Process', 'Date']).size()
axis = grouped.plot(kind='bar')
plt.show()

Where plt is the usual import matplotlib.pyplot as plt .

You may need to fix the label at the bottom of each bar, if they are too large.

Your error

Your error comes from the fact that grouped is a Series and not a DataFrame , so you cannot iterate over it that way. Iterating over a series returns only the value. You should do:

for value in grouped:
    print(value)

to see the sizes, but you lose the index label. To get also the index label the solution is:

for name, row in zip(grouped.index, grouped):
    print(name)
    print(row)