Joel Underwood Joel Underwood - 1 year ago 369
Python Question

python pandas countifs using multiple criteria AND multiple data frames

Trying to create--in Python using multiple data frames--the equivalent of a countifs in Excel that would span multiple sheets.

I need a new column count of records on another data frame based on criteria from the current data frame.

See Excel impression of what I want to do in python, also here.

My goal?

  • Count exams on students data frame

  • by Student ID

  • with exam date >= enroll date

  • with exam date <= detail date

  • with exam grade >= 70

Basically the Excel equivalent would be...

=COUNTIFS(Summary!$B$1:$B$11, ">="&Detail!B2,
Summary!$B$1:$B$11, "<="&Detail!C2,
Summary!$C$1:$C$11, ">="&70,
Summary!$A$1:$A$11, "="&Detail!A2)

...where Summary is the primary data frame and Detail is the secondary data frame where I want to count records.

Found these answers in my research:

Not quite what I'm looking for, because they don't span multiple data frames. I was able to create a basic countifs for a singular data frame:

sum(1 for x in students['Student ID'] if x == 1)
sum(1 for x in exams['Exam Grade'] if x >= 70)

Answer Source

Basically what you'll want to do is set up two dataframes, say df1 for the "exams passed" information and df2 for the marks on each exam.

To get yourself started, you can read in your excel files like this:

df1 = pd.read_excel('filename1.xlsx')
df2 = pd.read_excel('filename2.xlsx')

Then for each row in df1 you want to segment df2 and get the length of the segmented dataframe.

First though you might want to make list of information for each row in df1, which could be done like this:

student_info = df1[['Student ID', 'Enrollment Date', 'Qualification Date']].values

Then you can iterate through the rows like this:

N_exams_passed = [] # Store counts for each student in a list

for s_id, s_enroll, s_qual in student_info:
    N_exams_passed.append(len(df2[(df2['Student ID']==s_id) &
                                  (df2['Exam Date']>=s_enroll) &
                                  (df2['Exam Date']<=s_qual) &

Then add/replace the column in df1:

df1['Exams Passed'] = N_exams_passed

In order to compare the dates properly you will need to convert them to datetime objects in each pandas dataframe, and I will leave this up to you. Hint: you can use the pd.to_datetime() function.

Recommended from our users: Dynamic Network Monitoring from WhatsUp Gold from IPSwitch. Free Download