Json writer.writerow（）不'；不要写入正确的列

我有三张

DynamoDB

表格。两个表的实例ID是应用程序的一部分，另一个是我所有帐户和标记元数据中所有实例的主表。我对这两个表进行了两次扫描，以获取实例ID，然后在主表中查询标记元数据。但是，当我尝试将其写入

CSV

文件时，我希望每个dynamo表的唯一输出有两个单独的标题部分。第一次迭代完成后，第二个文件写入到第一次迭代结束的最后一行，而不是从第二个头部分的顶部重新开始。下面是我的代码和一个输出示例，以说明这一点

code:

import boto3
import csv
import json 
from boto3.dynamodb.conditions import Key, Attr

dynamo = boto3.client('dynamodb')
dynamodb = boto3.resource('dynamodb')
s3 = boto3.resource('s3')

# Required resource and client calls
all_instances_table = dynamodb.Table('Master')
missing_response = dynamo.scan(TableName='T1')
installed_response = dynamo.scan(TableName='T2')

# Creates CSV DictWriter object and fieldnames 
with open('file.csv', 'w') as csvfile:
    fieldnames = ['Agent Not Installed', 'Not Installed Account', 'Not Installed Tags', 'Not Installed Environment', " ", 'Agent Installed', 'Installed Account', 'Installed Tags', 'Installed Environment']
    writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
    writer.writeheader()

    # Find instances IDs from the missing table in the master table to pull tag metadata 
    for instances in missing_response['Items']:
        instance_missing = instances['missing_instances']['S']
        #print("Missing:" + instance_missing)
        query_missing = all_instances_table.query(KeyConditionExpression=Key('ID').eq(instance_missing))

        for item_missing in query_missing['Items']:
            missing_id = item_missing['ID']
            missing_account = item_missing['Account']
            missing_tags = item_missing['Tags']
            missing_env = item_missing['Environment']
            # Write the data to the CSV file
            writer.writerow({'Agent Not Installed': missing_id, 'Not Installed Account': missing_account, 'Not Installed Tags': missing_tags, 'Not Installed Environment': missing_env})

    # Find instances IDs from the installed table in the master table to pull tag metadata
    for instances in installed_response['Items']:
        instance_installed = instances['installed_instances']['S']
        #print("Installed:" + instance_installed)
        query_installed = all_instances_table.query(KeyConditionExpression=Key('ID').eq(instance_installed))

        for item_installed in query_installed['Items']:
            installed_id = item_installed['ID']
            print(installed_id)
            installed_account = item_installed['Account']
            installed_tags = item_installed['Tags']
            installed_env = item_installed['Environment']

            # Write the data to the CSV file 
            writer.writerow({'Agent Installed': installed_id, 'Installed Account': installed_account, 'Installed Tags': installed_tags, 'Installed Environment': installed_env})

输出：

这就是文件中的列/行的外观。

我需要每个标题部分的所有输出都在同一行上

数据：

下面是两个表的示例

示例输出：

以下是

for

循环打印出来并附加到列表中的内容

缺失：

i-0xxxxxx 333333333 foo@bar.com int 
i-0yyyyyy 333333333 foo1@bar.com int

已安装：

i-0zzzzzz 44444444 foo2@bar.com int
i-0aaaaaa 44444444 foo3@bar.com int

---

您希望将相关行收集到一个列表中，以便在一行上写入，例如：

missing = [] # collection for missing_responses
installed = [] # collection for installed_responses

# Find instances IDs from the missing table in the master table to pull tag metadata 
for instances in missing_response['Items']:
    instance_missing = instances['missing_instances']['S']
    #print("Missing:" + instance_missing)
    query_missing = all_instances_table.query(KeyConditionExpression=Key('ID').eq(instance_missing))
    for item_missing in query_missing['Items']:
        missing_id = item_missing['ID']
        missing_account = item_missing['Account']
        missing_tags = item_missing['Tags']
        missing_env = item_missing['Environment']
        # Update first half of row with missing list
        missing.append(missing_id, missing_account, missing_tags, missing_env)

# Find instances IDs from the installed table in the master table to pull tag metadata
for instances in installed_response['Items']:
    instance_installed = instances['installed_instances']['S']
    #print("Installed:" + instance_installed)
    query_installed = all_instances_table.query(KeyConditionExpression=Key('ID').eq(instance_installed))

    for item_installed in query_installed['Items']:
        installed_id = item_installed['ID']
        print(installed_id)
        installed_account = item_installed['Account']
        installed_tags = item_installed['Tags']
        installed_env = item_installed['Environment']
        # update second half of row by updating installed list
        installed.append(installed_id, installed_account, installed_tags, installed_env)
# combine your two lists outside a loop
this_row = []
i = 0;
for m in missing:
    # iterate through the first half to concatenate with the second half
    this_row.append( m + installed[i] )
    i = i +1

# adding an empty column after the write operation, manually, is optional
# Write the data to the CSV file 
writer.writerow(this_row)

---

如果已安装的表和丢失的表在一个可关联的字段（如时间戳或帐户ID）上运行，这将起作用，您可以确保以相同的顺序连接行。数据示例对于真正回答这个问题非常有用。

您需要编写代码来组合数据，然后每行编写一次。每次调用

writeRow（）

都会生成一个新行，不会向现有行添加额外的列。您是说“正确的列”吗？您可能想用“correct”而不是“right”来更新您的问题，因为列被视为位置：左、右、中等@Smittles我已经更新了问题。我已经用我的表格截图和

for

循环的一些

输出数据更新了问题。我推荐的方法可以，但这似乎与你希望他们在同一排的原因不符。如果是我，我会有另一列，一个布尔值，叫做“Installed”，它的字段是Yes或No（或True/False或1/0）。然后就可以进行分类了。您的循环数据没有任何可关联性-帐户不同，实例不重叠，因此（在我看来）没有理由将它们存储在同一行上。为什么他们在同一排？