如何让Python字典转YAML时仅按\n拆分字符串？

问题核心在于两点：一是禁止PyYAML自动折行长字符串（默认宽度限制导致），二是让原字符串中的\n转换为实际换行而非保留转义标识。以下是两种实现方案：

方案1：使用YAML字面量样式（推荐，符合YAML规范） #

通过自定义字符串代表器，让包含换行的字符串以|字面量样式输出，同时禁用自动折行：

import yaml
from yaml.representer import SafeRepresenter

# 自定义代表器：含换行的字符串用字面量样式输出
def str_with_newline_representer(dumper, data):
    if '\n' in data:
        return dumper.represent_scalar('tag:yaml.org,2002:str', data, style='|')
    return dumper.represent_scalar('tag:yaml.org,2002:str', data)

# 注册代表器到safe_dump
SafeRepresenter.add_representer(str, str_with_newline_representer)

# 修正拼写错误后的字典数据
d = {'test': {'long_string':'this is a long string that does not successfully split when it sees the character \n which is an issue'}}

# 导出YAML
with open('./test.yaml', 'w+') as ff:
    yaml.safe_dump(d, ff, width=float('inf'))

输出的YAML内容：

test:
  long_string: |
    this is a long string that does not successfully split when it sees the character 
     which is an issue

方案2：强制双引号包裹并换行 #

如果需要保持双引号包裹的格式，自定义代表器直接输出带换行的双引号字符串：

import yaml
from yaml.representer import SafeRepresenter

def quoted_multiline_str_representer(dumper, data):
    if '\n' in data:
        # 拆分字符串并处理换行后的缩进对齐
        lines = data.split('\n')
        formatted = [f'"{lines[0]}'] + [' ' * 16 + line for line in lines[1:]]
        formatted[-1] += '"'
        return dumper.represent_scalar('tag:yaml.org,2002:str', '\n'.join(formatted), style='')
    return dumper.represent_scalar('tag:yaml.org,2002:str', data, style='"')

SafeRepresenter.add_representer(str, quoted_multiline_str_representer)

d = {'test': {'long_string':'this is a long string that does not successfully split when it sees the character \n which is an issue'}}

with open('./test.yaml', 'w+') as ff:
    yaml.safe_dump(d, ff, width=float('inf'))

输出的YAML内容：

test:
  long_string: "this is a long string that does not successfully split when it sees the character 
                which is an issue"

关键参数说明 #

• width=float('inf')：禁用PyYAML默认的80字符宽度限制，避免长文本被自动折行
• 自定义代表器：控制字符串输出样式，将原字符串中的\n转换为YAML中的实际换行，而非转义字符

内容的提问来源于stack exchange，提问作者Kspr