Modularize python code into reusable functions

Question

I have the below code snippet that loops over all the files in a directory and populates the pandas dataframes. I am new to Python, so just wondering whether any part of this code snippet can be modularized (put it in a function) so that they can be reusable.

df_precision_list_a = []
df_precision_list_b = []
df_precision_list_c = []

##Reading a list of files from a directory

for file in allfiles:
  print(file)
  xls = pd.ExcelFile(file)
  df_sheet1 = pd.read_excel(xls, 'sheet1')
  df_overall_data = pd.read_excel(xls, 'overall_data')
  list_a = []
  list_b = []
  list_c = []

  ##Populating year to all the lists
  year = df_overall_data ['date'].iloc[0].year
  list_a.append(year)
  list_b.append(year)
  list_c.append(year)

  ##Populating month to all the lists
  month = df_overall_data ['date'].iloc[0].month
  list_a .append(month)
  list_b.append(month)
  list_c.append(month)

  ##Populating precision value to all the lists 
  precision_actual_a = df_sheet1.loc[df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
  list_a.append(precision_actual_a)
  precision_actual_b = df_sheet1.loc[df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
  list_b.append(precision_actual_b)
  precision_actual_c = df_sheet1.loc[df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]
  list_c.append(precision_actual_c)

  ##Populating Q2 values(fixed values) to all lists
  list_a.append(77)
  list_b.append(65)
  list_c.append(72)

  ##Adding lists (list_a,list_b,list_c) to respective precision lists
  df_precision_list_a.append(list_a)
  df_precision_list_b.append(list_b)
  df_precision_list_c.append(list_c)

##Populating precision dataframes using their respective lists
df_precision_a=pd.DataFrame(df_precision_list_a,columns=['Year', 'Month', 'Q1', 'Q2'])
df_precision_b=pd.DataFrame(df_precision_list_b,columns=['Year', 'Month', 'Q1', 'Q2'])
df_precision_c=pd.DataFrame(df_precision_list_c,columns=['Year', 'Month', 'Q1', 'Q2'])

You're doing pretty close to the same thing for 3 pieces of information. Identify what is done for all 3 (goes in a function) and what is different (becomes argument for function) and you should be able to set this up to loop and call 1 function 3 times to get the same result. — gallen
– gallen, Commented Jul 15, 2020 at 22:50
Make a generic class that contains all the attributes or methods that you want to instantiate or call with different inputs. — Data_Is_Everything
– Data_Is_Everything, Commented Jul 15, 2020 at 22:53

Partha Mandal · Accepted Answer · 2020-07-16 01:05:39Z

1

Only a quick tweak, but, this should make the flow better:

def main_func(file):
  print(file)
  xls = pd.ExcelFile(file)
  df_sheet1 = pd.read_excel(xls, 'sheet1')
  df_overall_data = pd.read_excel(xls, 'overall_data')
  list_a = []
  list_b = []
  list_c = []

  def append_to_list(inp_lst):
        nonlocal list_a, list_b, list_c
        list_a.append(inp_lst[0])
        list_b.append(inp_lst[1])
        list_c.append(inp_lst[2])   

  ##Populating year to all the lists
  year = df_overall_data ['date'].iloc[0].year
  append_to_list([year]*3)

  ##Populating month to all the lists
  month = df_overall_data ['date'].iloc[0].month
  append_to_list([month]*3)

  ##Populating precision value to all the lists 
  precision_actual_a = df_sheet1.loc[df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
  precision_actual_b = df_sheet1.loc[df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
  precision_actual_c = df_sheet1.loc[df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]
  append_to_list([precision_actual_a,precision_actual_b,precision_actual_c])

  ##Populating Q2 values(fixed values) to all lists
  append_to_list([77,65,72])

  # ... some more stuff
  # ... returns the lists

and now call this function with in the for loop you had:

for file in allfiles:
    ## prep the main lists ...
    main_func(file)
    ## append and create df

edited Jul 16, 2020 at 1:05

answered Jul 16, 2020 at 0:25

Partha Mandal

1,45111 silver badges16 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

user3447653 Over a year ago

Thank you. I get the error "No binding for nonlocal 'list_a'"

Diego Suarez · Accepted Answer · 2020-07-16 03:04:51Z

You should be able to do it in a nice clean way, by using an object and then you can call any method in any order that you want, and the object itself will keep track of the lists and attribute values.

I included an initialize_reset method to initialize or clear all lists and attribute values so that you are able to do calculation repeatedly with the same object instance.

Note: also this object is not tested with actual data, it runs without error, so if you test it with data you might have to customize it even further, but with this template you should be able to get it running.

class PandasFrames:

    def __init__(self, directory_path):
        self.allfiles = glob.glob(directory_path + '/*')
        self.initialize_reset_values()

    def initialize_reset_values(self):
        self.df_precision_list_a = []
        self.df_precision_list_b = []
        self.df_precision_list_c = []
        self.list_a = []
        self.list_b = []
        self.list_c = []
        self.df_sheet1 = None
        self.df_overall_data = None
        self.df_precision_a = None
        self.df_precision_b = None
        self.df_precision_c = None

    def read_directory_files(self):
        ''' Reading a list of files from a directory
        '''
        for file in self.allfiles:
            print(file)
            xls = pd.ExcelFile(file)
            self.df_sheet1 = pd.read_excel(xls, 'sheet1')
            self.df_overall_data = pd.read_excel(xls, 'overall_data')

    def populate_year(self):
        '''Populating year to all the lists
        '''
        if self.df_overall_data:
            year = self.df_overall_data['date'].iloc[0].year
            self.list_a.append(year)
            self.list_b.append(year)
            self.list_c.append(year)

    def populate_month(self):
        '''Populating month to all the lists
        '''
        if self.df_overall_data:
            month = self.df_overall_data['date'].iloc[0].month
            self.list_a.append(month)
            self.list_b.append(month)
            self.list_c.append(month)

    def populate_precision_value(self):
        '''Populating precision value to 
           all the lists
        '''
        if self.df_sheet1:
            precision_actual_a = self.df_sheet1.loc[self.df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
            self.list_a.append(precision_actual_a)
            precision_actual_b = self.df_sheet1.loc[self.df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
            self.list_b.append(precision_actual_b)
            precision_actual_c = self.df_sheet1.loc[self.df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]
            self.list_c.append(precision_actual_c)

    def populate_q2(self, values=[]):
        '''Populating Q2 values(fixed values) 
           to all lists
        '''
        if len(values) == 3:
            value1, value2, value3 = values
            self.list_a.append(value1)
            self.list_b.append(value2)
            self.list_c.append(value3)

    def add_lists(self):
        '''Adding lists (list_a,list_b,list_c) 
           to respective precision lists
        '''
        self.df_precision_list_a.append(self.list_a)
        self.df_precision_list_b.append(self.list_b)
        self.df_precision_list_c.append(self.list_c)

    def populate_precision_dataframes(self):
        '''Populating precision dataframes 
           using their respective lists
        '''
        self.df_precision_a=pd.DataFrame(self.df_precision_list_a,columns=['Year', 'Month', 'Q1', 'Q2'])
        self.df_precision_b=pd.DataFrame(self.df_precision_list_b,columns=['Year', 'Month', 'Q1', 'Q2'])
        self.df_precision_c=pd.DataFrame(self.df_precision_list_c,columns=['Year', 'Month', 'Q1', 'Q2'])

Ted Klein Bergman · Accepted Answer · 2020-07-16 01:25:36Z

You don't have to do things in that order. If you instead to first read all data you need, you'll see another pattern emerge:

for file in allfiles:
    print(file)
    xls = pd.ExcelFile(file)
    df_sheet1 = pd.read_excel(xls, 'sheet1')
    df_overall_data = pd.read_excel(xls, 'overall_data')
    
    year = df_overall_data ['date'].iloc[0].year
    month = df_overall_data ['date'].iloc[0].month
    precision_actual_a = df_sheet1.loc[df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
    precision_actual_b = df_sheet1.loc[df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
    precision_actual_c = df_sheet1.loc[df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]
    
    list_a = []
    list_b = []
    list_c = []
    
    ##Populating year to all the lists
    list_a.append(year)
    list_b.append(year)
    list_c.append(year)
    
    ##Populating month to all the lists
    list_a .append(month)
    list_b.append(month)
    list_c.append(month)
    
    ##Populating precision value to all the lists 
    list_a.append(precision_actual_a)
    list_b.append(precision_actual_b)
    list_c.append(precision_actual_c)
    
    ##Populating Q2 values(fixed values) to all lists
    list_a.append(77)
    list_b.append(65)
    list_c.append(72)
    
    ##Adding lists (list_a,list_b,list_c) to respective precision lists
    df_precision_list_a.append(list_a)
    df_precision_list_b.append(list_b)
    df_precision_list_c.append(list_c)

You might notice that all the append's are unnecessary. We can add all values directly into the lists:

for file in allfiles:
    print(file)
    xls = pd.ExcelFile(file)
    df_sheet1 = pd.read_excel(xls, 'sheet1')
    df_overall_data = pd.read_excel(xls, 'overall_data')

    year = df_overall_data ['date'].iloc[0].year
    month = df_overall_data ['date'].iloc[0].month
    precision_actual_a = df_sheet1.loc[df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
    precision_actual_b = df_sheet1.loc[df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
    precision_actual_c = df_sheet1.loc[df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]

    list_a = [year, month, precision_actual_a, 77]
    list_b = [year, month, precision_actual_b, 65]
    list_c = [year, month, precision_actual_c, 72]

    ##Adding lists (list_a,list_b,list_c) to respective precision lists
    df_precision_list_a.append(list_a)
    df_precision_list_b.append(list_b)
    df_precision_list_c.append(list_c)

You might notice now that the temporary lists are unnecessary too:

for file in allfiles:
    print(file)
    xls = pd.ExcelFile(file)
    df_sheet1 = pd.read_excel(xls, 'sheet1')
    df_overall_data = pd.read_excel(xls, 'overall_data')

    year = df_overall_data ['date'].iloc[0].year
    month = df_overall_data ['date'].iloc[0].month
    precision_actual_a = df_sheet1.loc[df_sheet1['Region'] == 'AA', 'Precision'].iloc[0]
    precision_actual_b = df_sheet1.loc[df_sheet1['Region'] == 'BB', 'Precision'].iloc[0]
    precision_actual_c = df_sheet1.loc[df_sheet1['Region'] == 'CC', 'Precision'].iloc[0]

    df_precision_list_a.append([year, month, precision_actual_a, 77])
    df_precision_list_b.append([year, month, precision_actual_b, 65])
    df_precision_list_c.append([year, month, precision_actual_c, 72])

And now we have small enough code to be put in its own function. You could reduce the code a bit more and put it in a function like this:

def create_precision_lists(files):
    df_precision_lists = [[], [], []]
    
    for file in files:
        xls = pd.ExcelFile(file)
        df_sheet1 = pd.read_excel(xls, 'sheet1')
        df_overall_data = pd.read_excel(xls, 'overall_data')
    
        year = df_overall_data ['date'].iloc[0].year
        month = df_overall_data ['date'].iloc[0].month
        
        for i, region in ('AA', 'BB', 'CC'):
            precision_actual = df_sheet1.loc[df_sheet1['Region'] == region, 'Precision'].iloc[0]
            df_precision_lists[i] = [year, month, precision_actual, 77]
    
    df_precision = [pd.DataFrame(precision_list, columns=['Year', 'Month', 'Q1', 'Q2']) for precision_list in df_precision_lists]
    return df_precision

It's not a great modular function though. To make it modular, you have to remove all the hardcoded values and you shouldn't rely on that specific keys and attributes exist. However, it's hard to do anything about it without knowing the full context of what you're trying to do and what your hard requirements and constraints are.

Collectives™ on Stack Overflow

Modularize python code into reusable functions

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related