pythonpandasdataframenumba

Why do I get "NameError: name 'df' is not defined"?

I don't know why "df" is not defined because i already assigned df a value. I read some question and answers online regarding this matter but none seem to help me solve my problem. If i remove the code to use CUDA, it runs perfectly but the project i'm working on requires for me to use CUDA.

@nb.jit(nopython=True)
def func_nb():
    data = pd.read_csv('dummycsv.csv')
    df = DataFrame(data)
    df['Tag'] = df['Tag'].map(lambda x: re.sub(r'\W+', '_', x))
tempdata = pd.DataFrame(columns=['Time','Value'])

taglist = pd.read_excel('dummyexcel.xlsx')
dt = DataFrame(taglist, columns=['Tag'])
dt['Tag'] = dt['Tag'].map(lambda x: re.sub(r'\W+', '_', x))


for i,row in dt.iterrows(): 

   tempdata = tempdata[0:0]      
   for index,row in df.iterrows():

       tag = dt.iloc[i]['Tag'] 
       x = df.iloc[index]['Tag']
       val = df.iloc[index]

Expected results are new files being created as a result from reading both excel and csv file and filtering the data in them.

Any help would be greatly appreciated!

Solution 
  • def func_nb():
   data = pd.read_csv('dummycsv.csv')
   df = DataFrame(data)
   df['Tag'] = df['Tag'].map(lambda x: re.sub(r'\W+', '_', x))
   return df

df = func_nb()

    After that you can use your dataframe, because it now exists outside of the your function.