Python Forum

Can someone please help me parse this json object?

{
    "12345": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
		},
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
		}
	},
    "12346": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
		},
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
		}
	}
}

The first number is a product ID, to which there are many.
The second is a date, to which there will be thousands.
Then I want to grab the key and value within each date.

I am looking to get Product ID, Date, A, B, and C into python variables.

I am just really struggling with this.

you can easily iterate over dict of dicts of dicts. the real question is what you are going to do with the data

(Jan-25-2021, 01:12 PM)buran Wrote: [ -> ]you can easily iterate over dict of dicts of dicts. the real question is what you are going to do with the data

I tried this...

for product in json:
    print(product)
    for date in product:
        print(date)

This shows the product ID as I expected.
But the date under products, doesn't

I'm probably going to stuck in into a DB for now.

You can loop.

data = {
    "12345": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
        },
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
        }
    },
    "12346": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
        },
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
        }
    }
}


for customer_id, dates in data.items():
    for date, values in dates.items():
        for key, value in values.items():
            print(customer_id, date, key, value, sep=", ")
            # here is maybe one level too deep
            # depends on your task

Or you can access data directly with subscription:

value = data["12345"]["2021-01-20"]["c"]

don't use json as name, it's a built-in module and you don't want to override it

data = {
    "12345": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
        },
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
        }
    },
    "12346": {
        "2021-01-20": {
            "a": 1,
            "b": 3,
            "c": 0
        },
        "2021-01-21": {
            "a": 1,
            "b": 3,
            "c": 0
        }
    }
}

for product, dates in data.items():
    for adate, values in dates.items():
        print(f"{product},{adate},{values['a']},{values['b']},{values['c']}")

Output:
12345,2021-01-20,1,3,0
12345,2021-01-21,1,3,0
12346,2021-01-20,1,3,0
12346,2021-01-21,1,3,0

(Jan-25-2021, 01:44 PM)buran Wrote: [ -> ]don't use json as name, it's a built-in module and you don't want to override it

Here a small example, what happens:

import json

print(json) # <-- shows information where the module is located

json = [1, 2, 3] # <-- assign a list to the name `json`
print(json) # <-- Print the object

# now trying to use a function from json module
json.dumps([1,2,3]) # <-- AttributeError

Output:
<module 'json' from 'c:\\users\\andre\\appdata\\local\\programs\\python\\python39\\lib\\json\\__init__.py'>
[1, 2, 3]

Error:
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-1-e9bf17c080eb> in <module>
      5 json = [1, 2, 3]
      6 print(json)
----> 7 json.dumps([1,2,3]) # <-- AttributeError

AttributeError: 'list' object has no attribute 'dumps'

If you import json, then the module json is loaded and assigned to the name json on module level.

The exception happens because the name json was assigned to something else.
In this case, the list has no attribute named dumps. dumps is a function from the json module, but not from a list. And this happens if you don't take care of names and assigning names which already exists.

You can even do this with built-in functions, which leads into unexpected behavior.

To avoid this, don't reassign objects to the name of imported modules and built-in function.
List of built-in functions: https://docs.python.org/3/library/functions.html
If you use an IDE, often these words are highlighted, if you assign accidentally a new object.
For example, sum is also a built-in, but often in code the programmers assign a value to sum and later they want to use the built-in sum function and BOOM.