How to append a new value to the same key in the list of dictionaries?

Ira :

I have a list of test-failures as shown below -

all_failures = [
            'test1/path/to/test1/log/failure_reason1',
            'test1/path/to/test1/log/failure_reason2',
            'test2/path/to/test2/log/failure_reason1',
            'test2/path/to/test2/log/failure_reason2',
            'test3/path/to/test3/log/failure_reason1',
            'test4/path/to/test4/log/failure_reason1',
        ]

I am trying to construct a JSON like object by parsing each failure in the list. So far, i have tried to write the following code -

for failure in all_failures:
    data = failure.split('/',1)
    test = data[0]
    failure_details_dict[test] = []

    data = '/' + data[1]
    data = data.rsplit('/', 1)

    test_details_dict['path'] = data[0] + '/'
    test_details_dict['reason'] = data[1]

    failure_details_dict[test].append(test_details_dict)

    test_details_dict = {}  

for key,value in failure_details_dict.items():
    print(key)
    print(value)
    print()

The output i am getting is -

test4
[{'reason': 'failure_reason1', 'path': '/path/to/test4/log/'}]

test3
[{'reason': 'failure_reason1', 'path': '/path/to/test3/log/'}]

test1
[{'reason': 'failure_reason2', 'path': '/path/to/test1/log/'}]

test2
[{'reason': 'failure_reason2', 'path': '/path/to/test2/log/'}]

whereas, the expected output is -

{
    "test1": [
                {
                    "path": "/path/to/test1/log/",
                    "reason": "failure_reason1" 
                },
                {
                    "path": "/path/to/test1/log/",
                    "reason": "failure_reason2"     
                }

            ],
    "test2": [
                {
                    "path": "/path/to/test2/log/",
                    "reason": "failure_reason1" 
                },
                {
                    "path": "/path/to/test2/log/",
                    "reason": "failure_reason2"     
                }
            ],
    "test3": [
                {
                    "path": "/path/to/test3/log/",
                    "reason": "failure_reason1" 
                },
            ],
    "test4": [
                {
                    "path": "/path/to/test4/log/",
                    "reason": "reason1" 
                },
            ]
}

As we can see, I have not been able to add the second path and reason for failure to the same key. Example - test1 and test2 have two reasons for failure.

Can someone please help to understand what i am missing? Thank you!

Boseong Choi :

Reason

You are overwriting to failure_details_dict[test] for every each loop.

---

Solution

You should set list to it only once.
You have several options to do it.

• Non-pythonic way(NOT RECOMMENDED)

if test not in failure_details_dict:
    failure_details_dict[test] = []

• Replace assignment to dict.setdefault call. This way doesn't affect other interactions with failure_details_dict

failure_details_dict.setdefault(test, [])  # instead of failure_details_dict[test] = []

• Use collections.defaultdict instead of dict. This way will AFFECT other interactions with failure_detilas_dict.

from collections import defaultdict

failure_details_dict = defaultdict(list)  # instead of {}

---

Example

And I have refactored your code:

all_failures = [
    'test1/path/to/test1/log/failure_reason1',
    'test1/path/to/test1/log/failure_reason2',
    'test2/path/to/test2/log/failure_reason1',
    'test2/path/to/test2/log/failure_reason2',
    'test3/path/to/test3/log/failure_reason1',
    'test4/path/to/test4/log/failure_reason1',
]

failure_details_dict = {}

for failure in all_failures:
    key, *paths, reason = failure.split('/')
    failure_details_dict.setdefault(key, []).append({
        'path': f"/{'/'.join(paths)}/",
        'reason': reason,
    })

for key, value in failure_details_dict.items():
    print(key)
    print(value)
    print()

---

Conclusion

• If you want a simple change, use dict.setdefault method.
• If you have multiple accesses to failure_details_dict and you want default value for each access, use collection.defaultdict class.

---

Extra

How can we modify the code so that 'path' key is copied only once and only multiple dictionaries with 'reason' key is created? In general, what would be the best way to store the data in JSON format?

You can reformat your JSON like:

{
  "test1": {
    "path": "/path/to/test1/log/",
    "reason": [
      "failure_reason1",
      "failure_reason2"
    ]
  },
  "test2": {
    "path": "/path/to/test2/log/",
    "reason": [
      "failure_reason1",
      "failure_reason2"
    ]
  },
  "test3": {
    "path": "/path/to/test3/log/",
    "reason": [
      "failure_reason1"
    ]
  },
  "test4": {
    "path": "/path/to/test4/log/",
    "reason": [
      "reason1"
    ]
  }
}

From code:

all_failures = [
    'test1/path/to/test1/log/failure_reason1',
    'test1/path/to/test1/log/failure_reason2',
    'test2/path/to/test2/log/failure_reason1',
    'test2/path/to/test2/log/failure_reason2',
    'test3/path/to/test3/log/failure_reason1',
    'test4/path/to/test4/log/failure_reason1',
]

failure_details_dict = {}

for failure in all_failures:
    key, *paths, reason = failure.split('/')
    failure_details_dict.setdefault(key, {
        'path': f"/{'/'.join(paths)}/",
        'reason': [],
    })['reason'].append(reason)

for key, value in failure_details_dict.items():
    print(key)
    print(value)
    print()