如何遍历Pandas DataFrame按花卉分组查找最小距离？

核心思路 #

不用手动遍历每一行，利用Pandas的groupby按花卉名分组，结合geopy批量计算每个种植地到目标地点（如芝加哥）的距离，再对每组取最小值，高效处理3000+行数据。

步骤实现 #

1. 安装依赖 #

先确保安装geopy库：

pip install geopy

2. 完整代码示例 #

import pandas as pd
from geopy.geocoders import Nominatim
from geopy.distance import geodesic

# 初始化地理编码器，user_agent设为你的应用标识（比如个人邮箱）
geolocator = Nominatim(user_agent="flower_distance_calculator")

# 预先获取目标地点（芝加哥）的经纬度，避免重复请求
target_location = geolocator.geocode("Chicago")
target_coords = (target_location.latitude, target_location.longitude)

# 定义计算单地点到芝加哥的英里距离函数
def calculate_distance(place):
    try:
        location = geolocator.geocode(place)
        if location:
            place_coords = (location.latitude, location.longitude)
            return geodesic(target_coords, place_coords).miles
        return float('inf')  # 找不到地点时设为无穷大，不影响取最小值
    except:
        return float('inf')

# 替换成你的实际DataFrame数据
data = {
    "Name": ["Geranium", "Geranium", "Geranium", "Daffodil", "Daffodil", "Daffodil", "Daffodil", "Carnations", "Carnations"],
    "Place": ["Paris", "Amsterdam", "Copenhagen", "Washington", "Helsinki", "Tokyo", "Sydney", "Kiambu", "Athens"]
}
df = pd.DataFrame(data)

# 按花卉分组，计算每组的最小距离
min_distances = df.groupby("Name")["Place"].apply(lambda x: x.apply(calculate_distance).min()).reset_index(name="Min Distance (miles)")

# 格式化输出为要求的列表形式
for _, row in min_distances.iterrows():
    print(f"- {row['Name']} - {row['Min Distance (miles)']:.2f} miles")

3. 针对3000+行数据的优化建议 #

• 缓存地理编码结果：如果DataFrame中有重复的Place，用字典缓存已编码的经纬度，避免重复调用Nominatim接口，大幅提升速度：

  coords_cache = {}
  def calculate_distance(place):
      if place in coords_cache:
          place_coords = coords_cache[place]
      else:
          try:
              location = geolocator.geocode(place)
              if location:
                  place_coords = (location.latitude, location.longitude)
                  coords_cache[place] = place_coords
              else:
                  return float('inf')
          except:
              return float('inf')
      return geodesic(target_coords, place_coords).miles
  

• 批量地理编码：如果数据量极大，可考虑使用批量地理编码工具，减少接口请求次数，进一步提升效率。

输出示例 #

运行代码后会得到类似结果：

• Geranium - 1732.40 miles
• Daffodil - 563.78 miles
• Carnations - 3943.21 miles

内容的提问来源于stack exchange，提问作者harjon