Python Pandas實(shí)現(xiàn)數(shù)據(jù)分組求平均值并填充nan的示例

factordatafillna = pd.DataFrame()
industrys = newfactordata1.industryName1.unique()
for ind in industrys:
  t = newfactordata1.industryName1 == ind
  a = newfactordata1[t].fillna(newfactordata1[t].mean())
  factordatafillna = factordatafillna.append(a)

df = pd.DataFrame({'code':[1,2,3,4,5,6,7,8],
          'value':[np.nan,5,7,8,9,10,11,12],          
          'value2':[5,np.nan,7,np.nan,9,10,11,12],
          'indstry':['農(nóng)業(yè)1','農(nóng)業(yè)1','農(nóng)業(yè)1','農(nóng)業(yè)2','農(nóng)業(yè)2','農(nóng)業(yè)4','農(nóng)業(yè)2','農(nóng)業(yè)3']},
          columns=['code','value','value2','indstry'],
          index=list('ABCDEFGH'))

# 只留下需要處理的列
cols = [col for col in df.columns if col not in['code','indstry']]
# 分組的列
gp_col = 'indstry'
# 查詢nan的列
df_na = df[cols].isna()
# 根據(jù)分組計(jì)算平均值
df_mean = df.groupby(gp_col)[cols].mean()

print(df)

# 依次處理每一列
for col in cols:
  na_series = df_na[col]
  names = list(df.loc[na_series,gp_col])   

  t = df_mean.loc[names,col]
  t.index = df.loc[na_series,col].index

  # 相同的index進(jìn)行賦值   
  df.loc[na_series,col] = t

print(df)

code value value2 indstry
A   1  NaN   5.0   農(nóng)業(yè)1
B   2  5.0   NaN   農(nóng)業(yè)1
C   3  7.0   7.0   農(nóng)業(yè)1
D   4  8.0   NaN   農(nóng)業(yè)2
E   5  9.0   9.0   農(nóng)業(yè)2
F   6  10.0  10.0   農(nóng)業(yè)4
G   7  11.0  11.0   農(nóng)業(yè)2
H   8  12.0  12.0   農(nóng)業(yè)3
  code value value2 indstry
A   1  6.0   5.0   農(nóng)業(yè)1
B   2  5.0   6.0   農(nóng)業(yè)1
C   3  7.0   7.0   農(nóng)業(yè)1
D   4  8.0  10.0   農(nóng)業(yè)2
E   5  9.0   9.0   農(nóng)業(yè)2
F   6  10.0  10.0   農(nóng)業(yè)4
G   7  11.0  11.0   農(nóng)業(yè)2
H   8  12.0  12.0   農(nóng)業(yè)3