python：从文件中读取json数据并追加更多数据

Question

I'm trying to read existing data from a json file and trying to append more data to the file using python (I'm a python newbie). Here is my existing data in data.json file which I read in my script:

{
    "Config1": {
        "TestCase1": {
            "Data1": 200,
            "Data2": 2715
        }
    },
    "Config2": {
        "TestCase1": {
            "Data1": 2710,
            "Data2": 2715
        }
    }
}

After reading I want to append TestCase2 data. This is what I'm doing:

with open("data.json") as json_file: #load existing data
    json_data = json.load(json_file)

test='TestCase2'
result=json_data
myConfigs = ['Config1','Config2']
for each, config in enumerate(myConfigs):
    result.update({config:{test:{'Data1':2600,'Data2':2900}}})
with open('data.json', 'a') as outfile:
    json.dump(result, outfile)

The new data in data.json is not valid as pointed by jsonLint . What am I doing wrong? Here is the new data

{
        "Config1": {
            "TestCase1": {
                "Data1": 200,
                "Data2": 2715
            }
        },
        "Config2": {
            "TestCase1": {
                "Data1": 2710,
                "Data2": 2715
            }
        }
    } {
        "Config1": {
            "TestCase2": {
                "Data1": 2600,
                "Data2": 2900
            }
        },
        "Config2": {
            "TestCase2": {
                "Data1": 2600,
                "Data2": 2900
            }
        }
    }

Answer 1

The problem is that you're appending the new JSON to the original JSON file here:

with open('data.json', 'a') as outfile:
    json.dump(result, outfile)

So you have two JSON objects in the same file as you can see:

...
            "Data2": 2715
        }
    }
} {  <--- original object ends here, new object starts here
    "Config1": {
...

JSONLint is expecting a single object, as will any JSON parser.

Answer 2

In addition to opening the file in the wrong mode (should be 'w'), you are also overwriting your old "config" trees by defining a new dict inline.

Instead of:

result.update({config:{test:{'Data1':2600,'Data2':2900}}})

Try this:

result[config][test] = {'Data1': 2600, 'Data2': 2900}

This should give you the result you are looking for with your example. It will let result['Config1']['TestCase1'] persist while you add TestCase2. You may also need to make sure that the config tree exists by setting result[config] to {} if it's None .

Answer 3

The main problem is that the dict1.update(dict2) method overwrites dict1 keys if the same exist in dict2 hence the second object in your file doesn't have key => TestCase1

Another problem is that (as pointed above) the file is opened in the wrong mode (Should be w ) as a appends to the json file

You could try this:

with open("data.json") as json_file:
    json_data = json.load(json_file)

test='TestCase2'
result=json_data
myConfigs = ['Config1','Config2']
for each, config in enumerate(myConfigs):
    result[config].update({test:{'Data1':2600,'Data2':2900}})
with open('data.json', 'w') as outfile:
    json.dump(result, outfile)

Just result[config].update(... instead of result.update({config:...