Question

我正在使用此dataframe（df1）：

df1 =  pd.read_csv('City_Zhvi_AllHomes.csv',header=None)

其内容包含以年 - 月（1-12）格式编写的列。现在，我必须创建一个新的dataframe，其中包含来自Df1的数据，分为4个季度。这就是我所做的：

Month = ['00','01','02','03','04','05','06','07','08','09','10','11','12']

Year=['00','01','02','03','04','05','06','07','08','09','10','11','12','13','14','15','16']


for i in range(0,15):

      for j in range(0,12):

          newdf = pd.DataFrame[df1.loc[2,'20'+Year[i]+'-'+Month[j+1]] + df1.loc[2,'20'+Year[i]+'-'+Month[j+2]]+ df1.loc[2,'20'+Year[i]+'-'+Month[j+3]],columns = ['Q1','Q2','Q3','Q4'] ] 
          j = j + 3


print newdf .

我已经专门为第二行做了这个，但我需要为整个数据集执行此操作

现在我收到了这个错误：

'type'对象没有属性' getitem '

以下是样本数据：

0             CountyName  SizeRank  2000-01  2000-02  2000-03  2000-04   
1                 Queens         1      NaN      NaN      NaN      NaN   
2            Los Angeles         2   204400   207000   209800   212300   
3                   Cook         3   136800   138300   140100   141900   
4           Philadelphia         4    52700    53100    53200    53400

newdf应该包含：

Q1                      
(204400+207000+209800)

同样Q2，Q3，Q4

我该如何处理？

Answer 1

我认为您可以使用axis=1列sum来汇总CountyName，但首先必须resample列SizeRank和#remove header=None for first row in csv as columns df = pd.read_csv('City_Zhvi_AllHomes.csv') df = df.set_index(['CountyName','SizeRank']) #with real data set index of all not dates columns #df = df.set_index(['RegionID', 'RegionName', 'State', 'Metro', 'CountyName', 'SizeRank']) df.columns = pd.to_datetime(df.columns).to_period('M') print (df) 2000-01 2000-02 2000-03 2000-04 CountyName SizeRank Queens 1 NaN NaN NaN NaN Los Angeles 2 204400.0 207000.0 209800.0 212300.0 Cook 3 136800.0 138300.0 140100.0 141900.0 Philadelphia 4 52700.0 53100.0 53200.0 53400.0 print (df.columns) PeriodIndex(['2000-01', '2000-02', '2000-03', '2000-04'], dtype='period[M]', freq='M') df = df.resample('Q', axis=1).sum() print (df) 2000Q1 2000Q2 CountyName SizeRank Queens 1 NaN NaN Los Angeles 2 621200.0 212300.0 Cook 3 415200.0 141900.0 Philadelphia 4 159000.0 53400.0然后首先转换列名称set_index然后转换to_datetime：

def bubble_sort_by(array)
  array.each_with_index do |outer, outer_index|
    array.each_with_index do |inner, inner_index|
      if yield(outer, inner) > 0
        if outer_index < inner_index
              p "outer after if: " + outer        # for debugging, indented for clarity
              p "inner after if: " + inner        # for debugging
          outer, inner = inner, outer
              p "outer after swap: " + outer      # for debugging
              p "inner after swap: " + inner      # for debugging
        end
      end
    end
  end
end

bubble_sort_by(["hi","hello","hey"]) do |left,right|
  left.length - right.length
end

从数据框访问项目并使用相同的方法创建一个新项目

1 个答案: