python - 将带有正确缩进的嵌套字典转储到 yaml

Question

I'm painfully close to automating the dump of a yml file that's created from a data frame for an automation task.

I have a function that is structured like the following:

def get_all_values(nested_dictionary):
    for key,value in nested_dictionary.items():
        model = {
           "models": [
                {
            "name": key,
              "columns": None
                }
            ]
        }            
        yield(model)
        for key,value in value.items():
                table = [
                    {
                       "name": key,
                            "tests": [            
                                "not_null",
                                "unique"            
                            ]
                       }
                ]
                yield(table)
    nested_dictionary = d1
    get_all_values(nested_dictionary)
    data = get_all_values(nested_dictionary)
    with open('data.yml', 'w') as outfile:
        with redirect_stdout(outfile):
            for i in data:
                ruamel.yaml.round_trip_dump(i,outfile, indent=5, block_seq_indent=2)

The dict it references is yielded as a generator. The dicts structure is:

    {'models': [{'name': 'budgets_sales', 'columns': None}]}
[{'name': 'budget_amt', 'tests': ['not_null', 'unique']}]
[{'name': 'budget_group', 'tests': ['not_null', 'unique']}]
[{'name': 'budget_name', 'tests': ['not_null', 'unique']}]
[{'name': 'budget_pk', 'tests': ['not_null', 'unique']}]
        

This works "well'...but the output is as follows:

models:
  -  name: budgets_sales
     columns:
  -  name: budget_amt
     tests:
       -  not_null
       -  unique
  -  name: budget_group
     tests:
       -  not_null
       -  unique
  -  name: budget_name
     tests:
       -  not_null
       -  unique

I require all values to keys in the dict to have an additional indentation. I cannot figure out how to make the values indented against the keys.

It would look like this if correct:

- name: budgets_sales
  columns:
      -  name: budget_amt
         tests:
            -  not_null
            -  unique
      -  name: budget_group
         tests:
            -  not_null
            -  unique
      -  name: budget_name
         tests:
            -  not_null
            -  unique
      -  name: budget_pk
         tests:
            -  not_null
            -  unique
      -  name: entry_type_code
         tests:
            -  not_null
            -  unique
      -  name: institution_fk
         tests:
            -  not_null
            -  unique

Could anyone provide an approach?

---

Thanks to Anthon, this is what I ended up using:

def get_all_values(nested_dictionary):
    res = [{"version":2},{"models":None}]
    for key,value in nested_dictionary.items():
        seq = []
        res.append([{"name": key, "columns": seq}])
        # for key1, value1 in value.items():  # not using value1
        for key1 in value.keys():
            elem = {"name": key1, "tests": ["not_null", "unique"]}
            seq.append(elem)
    return res

nested_dictionary = d1

get_all_values(nested_dictionary)

data = get_all_values(nested_dictionary)

    
with open('data.yml', 'w') as outfile:
    
    with redirect_stdout(outfile):
        
        for i in data:  
            
            yaml = ruamel.yaml.YAML()
            yaml.indent(mapping=5, sequence=5, offset=4)            
            yml.dump(i,outfile)

Answer 1

In your required output the value associated with key columns is a sequence. You only get that if your Python data structure is a list, so make sure you append your individual table entries to some variable.

I am guessing d1 based on your "incorrect" output:

import sys
import ruamel.yaml

d1 = dict(budgets_sales=dict(budget_amt=None, budget_group=None, budget_name=None, budget_pk=None))

def get_all_values(nested_dictionary):
    res = []
    for key,value in nested_dictionary.items():
        seq = []
        res.append({"name": key, "columns": seq})
        # for key1, value1 in value.items():  # not using value1
        for key1 in value.keys():
            elem = {"name": key, "tests": ["not_null", "unique"]}
            seq.append(elem)
    return res
    
data = get_all_values(d1)

yaml = ruamel.yaml.YAML()
yaml.indent(mapping=5, sequence=5, offset=3)
yaml.dump(data, sys.stdout)

which gives:

   - name: budgets_sales
     columns:
        - name: budgets_sales
          tests:
             - not_null
             - unique
        - name: budgets_sales
          tests:
             - not_null
             - unique
        - name: budgets_sales
          tests:
             - not_null
             - unique
        - name: budgets_sales
          tests:
             - not_null
             - unique

There are a few things you should take into account (apart from better formatting your code and data here on SO):

• the round_trip_dump function has been deprecated, don't use it in new code
• the recommended extension for files containing YAML documents has been .yaml , since at least Sep 2007
• don't write YAML files in multiple stages, create one complete data structure and dump it. If you want multiple YAML documents in one file, make a list of data structures and use the .dump_all() method.

If all else fails and you have valid handcrafted YAML that you want to generate as output, load that YAML (with YAML(typ='safe').load() and inspect the data structure in Python that you get.