使用Pandas DataFrame写入数据时出现'append'属性不存在错误

执行循环写入设备信息到DataFrame时，出现错误：

AttributeError: 'DataFrame' object has no attribute 'append'. Did you mean: '_append'?

可复现代码：

import pandas as pd
import os
import json

currDir = os.getcwd()
def parse_json_response():

    filename = "my_json_file.json"
    device_name = ["Trona", "Sheldon"]
    "creating dataframe to store result"
    column_names = ["DEVICE", "STATUS", "LAST UPDATED"]
    result_df = pd.DataFrame(columns=column_names)
    my_json_file = currDir + '/' + filename

    for i in range(len(device_name)):
        my_device_name = device_name[i]
        with open(my_json_file) as f:
            data = json.load(f)

        for devices in data:
            device_types = devices['device_types']
            if my_device_name in device_types['name']:
                if device_types['name'] == my_device_name:
                    device = devices['device_types']['name']
                    last_updated = devices['devices']['last_status_update']
                    device_status = devices['devices']['status']

                    result_df = result_df.append(
                      {'DEVICE': device, 'STATUS': device_status,
                     'LAST UPDATED': last_updated}, ignore_index=True)
    print(result_df)

parse_json_response()

JSON文件内容（保存为当前路径下的my_json_file.json）：

[{"devices": {"id": 34815, "last_status_update": "2023-05-25 07:56:49", "status": "idle" }, "device_types": {"name": "Trona"}}, {"devices": {"id": 34815, "last_status_update": "2023-05-25 07:56:49", "status": "idle" }, "device_types": {"name": "Sheldon"}}]

原因与解决方案 #

报错原因 #

Pandas 2.0及后续版本已正式移除DataFrame.append()方法，这是触发该错误的核心原因。

解决方案 #

推荐两种替代方式，同时优化原代码中的冗余操作：

方案1：改用pd.concat()合并DataFrame

将每次要添加的行转为单个DataFrame，再用pd.concat()合并到结果中：

import pandas as pd
import os
import json

currDir = os.getcwd()
def parse_json_response():

    filename = "my_json_file.json"
    device_name = ["Trona", "Sheldon"]
    column_names = ["DEVICE", "STATUS", "LAST UPDATED"]
    result_df = pd.DataFrame(columns=column_names)
    my_json_file = os.path.join(currDir, filename)  # 用os.path.join处理路径更安全

    # 只读取一次JSON文件，避免循环重复IO操作
    with open(my_json_file) as f:
        data = json.load(f)

    for my_device_name in device_name:  # 直接遍历列表，无需通过索引取值
        for devices in data:
            device_types = devices['device_types']
            # 去掉重复判断，直接判断名称相等即可
            if device_types['name'] == my_device_name:
                device = devices['device_types']['name']
                last_updated = devices['devices']['last_status_update']
                device_status = devices['devices']['status']

                # 用pd.concat替代append
                new_row = pd.DataFrame([{'DEVICE': device, 'STATUS': device_status,
                                        'LAST UPDATED': last_updated}])
                result_df = pd.concat([result_df, new_row], ignore_index=True)
    print(result_df)

parse_json_response()

方案2：先收集数据到列表，最后一次性生成DataFrame（效率更高）

这种方式避免多次合并DataFrame，性能更优，尤其数据量大时：

import pandas as pd
import os
import json

currDir = os.getcwd()
def parse_json_response():

    filename = "my_json_file.json"
    device_name = ["Trona", "Sheldon"]
    column_names = ["DEVICE", "STATUS", "LAST UPDATED"]
    my_json_file = os.path.join(currDir, filename)
    data_list = []  # 用列表存储所有行数据

    with open(my_json_file) as f:
        data = json.load(f)

    for my_device_name in device_name:
        for devices in data:
            device_types = devices['device_types']
            if device_types['name'] == my_device_name:
                data_list.append({
                    'DEVICE': devices['device_types']['name'],
                    'STATUS': devices['devices']['status'],
                    'LAST UPDATED': devices['devices']['last_status_update']
                })
    
    # 最后一次性生成DataFrame
    result_df = pd.DataFrame(data_list, columns=column_names)
    print(result_df)

parse_json_response()

额外优化点 #

1. 原代码在循环中重复读取JSON文件，改为只读取一次，减少IO开销
2. 直接遍历device_name列表，无需通过索引取值，代码更简洁
3. 去掉了重复的if my_device_name in device_types['name']判断，直接判断相等即可，逻辑更清晰

内容的提问来源于stack exchange，提问作者ilexcel