Preserve the original literal block format
No, it is not possible for standard Python YAML libraries like PyYAML to preserve the original literal block format (e.g., | or >) when you load a YAML file and then dump it back. This is because the parser converts the literal block content into a standard Python string, discarding the original formatting. When you dump the string back to YAML, the dumper uses its own rules to represent the string, which typically defaults to a quoted style or a folded block style, but not necessarily the original one.
The Problem in Detail
When a YAML parser reads a literal block, it processes the content and stores it in memory as a simple string. The information about whether the block was a literal block (|), a folded block (>), or a plain scalar is lost.
For example, when you parse this YAML:
text: |
This is
a literal block.
The parser sees the | and correctly interprets the value as a multi-line string: "This is\na literal block.\n". The fact that it was originally a literal block is not retained in the Python object.
When you dump this string back to YAML, the dumper's logic takes over. It will typically try to output the most readable representation of that string. If the string contains newlines, it might choose a folded block style (>) or a literal block style (|), but it has no way of knowing which style the original author used.
The Code Example
Here's a practical example using PyYAML to demonstrate this behavior:
import yaml
# Original YAML string with a literal block
yaml_string = """
document:
key: |
This is
a literal block
with preserved newlines.
"""
# 1. Load the YAML. The literal block is converted to a standard Python string.
data = yaml.safe_load(yaml_string)
# The loaded Python data is just a dictionary with a multi-line string
# {'document': {'key': 'This is\na literal block\nwith preserved newlines.\n'}}
print("Python data after loading:", repr(data['document']['key']))
# 2. Modify the data (or not).
# Let's say we just want to dump it back to a string.
dumped_yaml = yaml.dump(data, sort_keys=False)
# 3. Dump the data. The dumper makes its own choice for the format.
# It often uses a folded style by default, not the original literal style.
print("\nYAML after dumping:")
print(dumped_yaml)
The output of the dumped YAML might look something like this, using a folded style (>) or a quoted style, instead of the original literal block (|):
document:
key: >
This is a literal block with preserved newlines.
The Solution: Manual Intervention
To preserve the original formatting, you would need to write a custom YAML parser or dumper that tracks and handles these specific block types. However, this is a complex task and goes against the fundamental design of most YAML libraries, which are built to map YAML data to native Python types and vice versa.
The common workaround for this issue is to manually reconstruct the YAML output or to avoid parsing the files if you only need to modify parts that are not literal blocks.