从mongo导入具体数据到pandas dataframe

Question

I have a large amount of data in a collection in mongodb which I need to analyze, using pandas and pymongo in jupyter. I am trying to import specific data in a dataframe.

Sample data.

{
    "stored": "2022-04-xx",
    ...
    ...
    "completedQueues": [
        "STATEMENT_FORWARDING_QUEUE",
        "STATEMENT_PERSON_QUEUE",
        "STATEMENT_QUERYBUILDERCACHE_QUEUE"
    ],
    "activities": [
        "https://example.com
    ],
    "hash": "xxx",
    "agents": [
        "mailto:example@example.com"
    ],
    "statement": {                                  <=== I want to import the data from "statement"
        "authority": {
            "objectType": "Agent",
            "name": "xxx",
            "mbox": "mailto:example@example.com"
        },
        "stored": "2022-04-xxx",
        "context": {
            "platform": "Unknown",
            "extensions": {
                "http://example.com",
                "xxx.com": {
                    "user_agent": "xxx"
                },
                "http://example.com": ""
            }
        },
        "actor": {
            "objectType": "xxx",
            "name": "xxx",
            "mbox": "mailto:example@example.com"
        },
        "timestamp": "2022-04-xxx",
        "version": "1.0.0",
        "id": "xxx",
        "verb": {
            "id": "http://example.com",
            "display": {
                "en-US": "viewed"
            }
        },
        "object": {
            "objectType": "xxx",
            "id": "https://example.com",
            "definition": {
                "type": "http://example.com",
                "name": {
                    "en-US": ""
                },
                "description": {
                    "en-US": "Viewed"
                }
            }
        }
    },                                             <=== up to here
    "hasGeneratedId": true,
    ...
    ...
}

Notice that I am only interested in data nested under "statement", and not in any data containing the string, ie the "STATEMENT_FORWARDING_QUEUE" above it.

What I am trying to accomplish is import the data from "statement" (as indicated above) in a dataframe, and arrange them in a manner, like:

id	authority objectType	authority name	authority mbox	stored	context platform	context extensions	actor objectType	actor name	...
00	Agent	xxx	mailto	2022-	Unknown	http://1	xxx	xxx	...
01	Agent	yyy	mailto	2022-	Unknown	http://2	yyy	yyy	...

The idea is to be able to access any data like "authority name" or "actor objectType".

I have tried:

df = pd.DataFrame(list(collection.find(query)(filters)))
df = json_normalize(list(collection.find(query)(filters)))

with various queries, filter and slices, and also aggregate and map/reduce, but nothing produces the correct output.

I would also like to sort (newest to oldest) based on the "stored" field (sort('$natural',-1)?), and maybe apply limit(xx) to the dataframe as well.

Any ideas?

Thanks in advance.

Answer 1

Try this

df = json_normalize(list(
    collection.aggregate([
        {
            "$match": query
        },
        {
            "$replaceRoot": {
                "newRoot": "$statement"
            }
        }
    ])
)

Answer 2

Thanks for the answer, @pavel. It is spot on and pretty much solves the problem.

I also added sorting and limit, so if anyone is interested, the final code looks like this:

df = json_normalize(list(
  statements_coll.aggregate([
    {
        "$match": query
    },
    {
        "$replaceRoot": {
            "newRoot": "$statement"
        }
    },
    { 
        "$sort": { 
            "stored": -1 
        }
    },
    {
        "$limit": 10 
    }
  ]) 
))