Python ruamel.yaml 代码在控制台中有效，但在 function (spyder) 中无效

Question

I am working in Spyder, Python version 3.9. I am running into an issue where a bit of code I have written in a function isn't working, but when running the code in the console, I get the desired outcome.

I am trying to write a function which takes items from a list and writes them to a YAML file. When the code is in a function, I will get an error. Below is a sample of what I would type in the console to try and run. I copied over the function that I had written in the code block into the console and tried running the below.

import ruaml.yaml
import pandas as pd

def getTable(tableDir):
    rawData = pd.read_csv(tableDir)
    tableData = pd.DataFrame(rawData)
    return tableData

def createYamlFile(tableDataLine, yamlFileName):
    yamlString = """

    """

    for i in range(0, len(tableDataLine), 1):
        yamlString = "{}{}\n".format(yamlString, tableDataLine[i]

    yaml = ruamel.yaml.YAML()
    yaml.preserve_quotes = True
    yamlContent = ruamel.yaml.safe_load(yamlString)

    with open(yamlFileName, 'w') as file:
        yaml.dump(yamlContent, file)

table = getTable(tableDir)
createYamlFile(table.iloc[0])

The desired outcome would look something like the following:

---
#0: Test0
1: Test1
2: Test2
#3: Test3
4: Test4

However, when I run the first code block in the console as-is, I get a ComposerError: expected a single document in the stream .

When I run the exact same code in the console, outside of the function though eg

table = getTable(tableDir)

yamlString = """

"""

for i in range(0, len(table.iloc[0]), 1):
    yamlString = "{}{}\n".format(yamlString, table.iloc[0][i])

yaml = ruamel.yaml.YAML()
yaml.preserve_quotes = True
yamlContent = ruamel.yaml.safe_load(yamlString)

with open(yamlFileName, 'w') as file:
    yaml.dump(yamlContent, file)

I am able to create my file just fine. The comments don't show up but this is a secondary problem to me. What is happening? How come the exact same code in the console works when it's not in a function?

Answer 1

When you run this code:

yaml_str = """\

"""

def fun():
    yaml_str = """

    """
    return yaml_str

print(repr(yaml_str))
print(repr(fun()))

it gives:

'\n'
'\n\n    '

So your console and your program don't start with the same string. Since your code "{}{}\n".format(yamlString, tableDataLine[i] essentially concatenates the table data a line at a time to that string, the first line gets indented, which is likely to cause an error.

If you want to specify YAML input code within a function (or any other indented environment) use either: yaml_str = "\n" or if you need readable multiline input use textwrap.dedent :

from textwrap import dedent

def fun():
    yaml_str = dedent("""\

    """)
    return yaml_str

print(repr(fun()))

which gives:

'\n'

I always use a backslash escape after the starting triple quotes to get rid of the leading newline. You can also have multeline strings within a function that continue in the first column, but I find that unreadable.

---

A far bigger problem is you mising the old API ( ruamel.yaml.safe_load() ) the new API (yaml = `ruamel.yaml.YAML()). That will give problems and you should not be using the old API for any new code (if possible notify the place where you got this broken code example from). Instead do use:

import ruamel.yaml
from textwrap import dedent

def createYamlFile(tableDataLine, yamlFileName):
    yamlString = """

    """

    for i in range(0, len(tableDataLine), 1):
        yamlString = "{}{}\n".format(yamlString, tableDataLine[i])  # < closing parenthesis missing in non-console example

    yaml = ruamel.yaml.YAML()
    yaml.preserve_quotes = True
    yamlContent = yaml.load(yamlString)  # the default YAML() instance returns a subclass of the SafeLoader

    with open(yamlFileName, 'wb') as file:  # ruamel.yaml defaults to UTF-8, so use a binary file
        yaml.dump(yamlContent, file)