0

I have written a function in python:

import pandas as pd
from pandas import DataFrame, read_csv
import numpy as np
import os

def textFileReader(text_file_location,delimiter,header,columns_needed,*agrs):
    """when file has header and all columns are required,first section will be executed"""
    """when file has header and selected columns are required,second will be executed"""
    """when file hasn't header, last section will be executed"""

    if ((header =="yes") and (columns_needed=="all")):
        file_name = pd.read_csv(text_file_location,sep=delimiter,header=0)
        col_names = [x.lower() for x in list(file_name.columns)]
        file_name.columns = col_names

    elif ((header=="yes") and (columns_needed=="columns")):
        file_name = pd.read_csv(text_file_location,sep=delimiter,header=0)
        file_name = file_name[columns]
        col_names = [x.lower() for x in list(file_name.columns)]
        file_name.columns = col_names
    else:
        file_name = pd.read_csv(text_file_location,sep=delimiter)

"""drop na's and duplicates from the data"""

    nas_dropped = file_name.dropna(axis = 0, how = 'all')
    duplicates_removed = file_name.drop_duplicates()

    """basic view of data"""

    print duplicates_removed.shape
    print duplicates_removed.dtypes
    print duplicates_removed.head()


   return duplicates_removed

I am using this function in other program abc.py

import pandas as pd
from pandas import DataFrame, read_csv
import numpy as np
import os
from textFileReader import textFileReader

path = os.path.join('/home','ubuntu','Nov_2014','amgen_content_recommendation_demo_data.tsv')
textFileReader.columns  = ['Phy_Id','Specialty','Age']
demo_data = textFileReader(path,"\t","yes","columns")

I am getting an error : 

NameError: global name 'columns' is not defined

I am passing the value of columns in elif section of my function. I am trying here to change values assigned to columns baesd on requirement while reading a text file. In another iteration, I may assign some other column names. Is ther any way to do this?

Though the error I am getting is same from many previous questions, what I am trying to do here is different. I have created a local variable "columns" in my main function and then assigning a list to that in my main program where the funcion has been called.

Pulkit Jha
  • 1,709
  • 3
  • 12
  • 18
  • 1
    Aren't global variables just global to a specific module, rather than crossing over into imported modules? [See this question](http://stackoverflow.com/questions/15959534/python-visibility-of-global-variables-from-imported-modules) – elParaguayo Dec 02 '14 at 11:53
  • And if you are using columns with in the function only, I don't think its necessary to declare as global. A default assignment would work. – Tanveer Alam Dec 02 '14 at 11:57
  • I have edited my code, but still getting the same error. – Pulkit Jha Dec 02 '14 at 14:57
  • Though the error I am getting is same, what I am trying to do here is different. I have created a local variable "columns" in my main function and then assigning a list to that in my main program where the funcion has been called. – Pulkit Jha Dec 03 '14 at 10:56

0 Answers0