Python：使用GroupBy对象上的过滤器过滤掉数据帧

数据集：

使用提供的字典，通过执行以下操作，根据通货膨胀调整新的价格：

 a   Make a function called inflation that inputs a dataframe.

 b   Make a new variable conversion. This is a column with the values of conversion_table that matches the column Year as the key.

 c   Remove any non-numerical characters in the column New_Price. Replace the New_Price with that change.

 d   Convert the column type New_Price into float. Replace the New_Price with that change.

 e   Multiply New_Price with conversion. Replace the New_Price with that change.

 f   Return the dataframe.
 
 Then, Print Year, New_Price, and New_Price_Adjusted.

我的代码如下：

# 1
bool_mean_less_than_average = df.groupby("Location")["Price"].mean() < df["Price"].mean()
filtered_less = df.groupby("Location")["Price"].mean()[bool_mean_less_than_average]

print(df.groupby("Location").filter(lambda group: group["Price"].mean() > df["Price"].mean())["Location"].unique())
# 2
bool_mean_higher_than_average = df.groupby("Location")["Price"].mean() > df["Price"].mean()
filtered_higher = df.groupby("Location")["Price"].mean()[bool_mean_higher_than_average]
print(df.groupby("Location").filter(lambda group: group["Price"].mean() < df["Price"].mean())["Location"].unique())
#3
conversion_table = {
    1998: 3.7327,
    1999: 3.2372,
    2000: 3.2216,
    2001: 3.1133,
    2002: 2.9603,
    2003: 2.8694,
    2004: 2.7662,
    2005: 2.6652,
    2006: 2.5246,
    2007: 2.3702,
    2008: 2.2461,
    2009: 2.0475,
    2010: 1.7809,
    2011: 1.6270,
    2012: 1.5278,
    2013: 1.3743,
    2014: 1.2594,
    2015: 1.1897,
    2016: 1.1189,
    2017: 1.0945,
    2018: 1.0524,
    2019: 1.0}

def inflation(df):
  conversion = conversion_table[df["Year"].values[0]]
  #df[df.New_Price.apply(lambda x: x.isnumeric())].set_index('New_Price')
  #df[pd.to_numeric(df['New_Price'], errors='coerce').isna()]
  df['New_Price'] = pd.to_numeric(df['Price'], errors='coerce')
  #df = df.dropna(subset=['New_Price']) 
  df["New_Price"] = df["New_Price"].apply(pd.to_numeric, downcast='float', errors='coerce')
  dep_plus_arr = df["New_Price"] 
  df["New_Price"] = (dep_plus_arr * conversion).round(1)
  
  return df
  #4
  df.groupby("Year",as_index=False).apply(inflation)
  print(df[["Year","New_Price"]])

#1
bool_mean_小于_average=df.groupby（“位置”）[“价格”].mean（）df[“价格”].mean（））[“位置”].unique（））
# 2
bool_mean_高于_average=df.groupby（“位置”）[“价格”].mean（）>df[“价格”].mean（）
过滤后的平均值=df.groupby（“位置”）[“价格”].mean（）[布尔平均值高于平均值]
打印（df.groupby（“位置”）.filter（lambda组：组[“价格”].mean（）

但是输出是如此奇怪：

不应该出现像NaN这样的输出。 如何解决这个问题