错误'AttributeError：'DataFrameGroupBy'对象没有属性'而数据框上的groupby功能

Question

I have a dataframe news_count . Here are its column names, from the output of news_count.columns.values :

 [('date', '') ('EBIX UW Equity', 'NEWS_SENTIMENT_DAILY_AVG') ('Date', '')
  ('day', '') ('month', '') ('year', '')]

I need to groupby by year and month and sum values of 'NEWS_SENTIMENT_DAILY_AVG' . Below is code I tried, but neither work:

Attempt 1

news_count.groupby(['year','month']).NEWS_SENTIMENT_DAILY_AVG.values.sum()

'AttributeError: 'DataFrameGroupBy' object has no attribute' 

Attempt 2

news_count.groupby(['year','month']).iloc[:,1].values.sum()

AttributeError: Cannot access callable attribute 'iloc' of 'DataFrameGroupBy' objects, try using the 'apply' method

Input data:

      ticker       date           EBIX UW Equity    month    year
      field             NEWS_SENTIMENT_DAILY_AVG
         0      2007-05-25                   0.3992      5       2007
         1      2007-11-06                   0.3936      11      2007 
         2      2007-11-07                   0.2039      11      2007
         3      2009-01-14                   0.2881       1      2014

Answer 1

extract required columns from dataframe in news_count_res variable and then apply aggregation function

news_count_res = news_count[['year','month','NEWS_SENTIMENT_DAILY_AVG']]
news_count_res.group(['year','month']).sum()

Answer 2

In following way we can apply aggregation functions on required columns without extracting columns as well

news_count.group(['year','month'])['NEWS_SENTIMENT_DAILY_AVG'].sum()

Answer 3

Thanks to answers so far (I've made comments there as I haven't got those solutions to work--maybe I'm not understanding something). In the meantime, I've also come up with another approach, which I still suspect isn't very Pythonic. It does get the job done and doesn't take too long for my purposes, but it would be great if I could figure out how to tweak the approaches suggested above to get them to work...any thoughts very welcome!

Here's what I've got:

    import pandas as pd
    import math
    y = ['Alex'] * 2321 + ['Doug'] * 34123  + ['Chuck'] * 2012 + ['Bob'] * 9281 
        z = ['xyz'] * len(y)
    df = pd.DataFrame({'persons': y, 'data' : z})
    percent = 10  #CHANGE AS NEEDED

    #add a 'helper'column with random numbers
    df['rand'] = np.random.random(df.shape[0])
    df = df.sample(frac=1)  #optional:  this shuffles data, just to show order doesn't matter

    #CREATE A HELPER LIST
    helper = pd.DataFrame(df.groupby('persons')['rand'].count()).reset_index().values.tolist()
    for row in helper:
        df_temp = df[df['persons'] == row[0]][['persons','rand']]
        lim = math.ceil(len(df_temp) * percent * 0.01)
        row.append(df_temp.nlargest(lim,'rand').iloc[-1][1])

    def flag(name,num):
        for row in helper:
            if row[0] == name:
                if num >= row[2]:
                    return 'yes'
                else:
                    return 'no'
    
    df['flag'] = df.apply(lambda x: flag(x['persons'], x['rand']), axis=1)

And to check the results:

piv = df.pivot_table(index="persons", columns="flag", values="data", aggfunc='count', fill_value=0)
piv = piv.apivend(piv.sum().rename('Total')).assign(Total=lambda x: x.sum(1))
piv['% selected'] = 100 * piv.yes/piv.Total
print(piv)

OUTPUT:
flag        no   yes  Total  % selected
persons                                
Alex      2088   233   2321   10.038776
Bob       8352   929   9281   10.009697
Chuck     1810   202   2012   10.039761
Doug     30710  3413  34123   10.002051
Total    42960  4777  47737   10.006913

Seems to work with different %s and different numbers of persons...but it would be nice to make it simpler, I think.