CODEWITHSUNDEEP
sqlarraysjsonsql-serversql-server-2012
Mike Eale
Mike Eale

Reputation: 121

Parse nested JSON in a column to Table in SQL Server

I have a table that only have one column and the records are in json format as shown below: enter image description here

The sample of each row is:

{
    "id": "51cf9ff0-0ed5-11eb-8887-53248e3b2424",
    "attributes": {
        "source": "Google",
        "medium": "cpc",
        "visit_route": [
            {
                "time_on_page": 5,
                "page_title": "Dedicated Servers"               
            },
            {
                "time_on_page": 1,
                "page_title": "Partner Programme"                
            }
        ],
        "keyword": null,
        "visit_length": 6,
        "started_at": "2020-10-15T10:56:31.51Z",
        "ga_client_ids": [
            "1213599109.1602733400"
        ],
       "lead_id": "597b4cd6-d8fb-11e6-adad-17d0cee77142_ayRRmwDGKhjjSgdcMGDMGf"
    }
}

The outcome should look like below:

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|id                                   |source |medium |visit_route                                                                                                  |Keyword|
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|51cf9ff0-0ed5-11eb-8887-53248e3b2424 |Google |cpc    |[{"time_on_page": 5,"page_title": "Dedicated Servers"},{"time_on_page": 1,"page_title": "Partner Programme"}]| Null  |

This sample is in each rows. I am new to parsing json in sql and have tried using the script below:

 select id ,
        attributes
from [StageDB].[dbo].[LeadFeeder_visits_json] 
  cross apply openjson(jsonObj)
  WITH (   
             id   nvarchar(100) ,  
                
             attributes nvarchar(max)   
                
        )

But the result I got as shown below:

enter image description here

I really need help.

Upvotes: 1

Views: 6214

Answers (3)

SteveC
SteveC

Reputation: 6015

An alternative to OPENJSON is to use JSON_VALUE and JSON_QUERY to pick out the pathways needed. The JSON functions can optionally specify whether or not a field is strictly required. For example, in the code the first JSON_VALUE specifies the 'id' to be 'strict'-ly present in the JSON. If the 'id' field is not present the JSON_VALUE function will return an error

JSON in a table

drop table if exists #json;
go
create table #json(
  jsonObj                   nvarchar(max) not null);
go

insert #json(jsonObj) values (N'{
    "id": "51cf9ff0-0ed5-11eb-8887-53248e3b2424",
    "attributes": {
        "source": "Google",
        "medium": "cpc",
        "visit_route": [
            {
                "time_on_page": 5,
                "page_title": "Dedicated Servers"               
            },
            {
                "time_on_page": 1,
                "page_title": "Partner Programme"                
            }
        ],
        "keyword": null,
        "visit_length": 6,
        "started_at": "2020-10-15T10:56:31.51Z",
        "ga_client_ids": [
            "1213599109.1602733400"
        ],
       "lead_id": "597b4cd6-d8fb-11e6-adad-17d0cee77142_ayRRmwDGKhjjSgdcMGDMGf"
    }
}');

Query

select json_value(jsonObj, N'strict $.id') id,
       json_value(jsonObj, N'$.attributes.source') [source],
       json_value(jsonObj, N'$.attributes.medium') [medium],
       json_query(jsonObj, N'$.attributes.visit_route') visit_route,
       json_value(jsonObj, N'$.attributes.keyword') keyword
from #json;

Output

id                                      source  medium  visit_route                                                                       keyword
51cf9ff0-0ed5-11eb-8887-53248e3b2424    Google  cpc     [{time_on_page": 5, "page_title": "Dedicated Servers" }, { "time_on_page": 1, "page_title": "Partner Programme" } ] NULL

Upvotes: 0

John Cappelletti
John Cappelletti

Reputation: 81930

If you'll have multiple sets of ATTRIBUTES, then GMB is correct with the second cross apply. If only one set, then you can get by with a single cross apply.

Example

Select B.*
 From  YourTable A
 Cross Apply OpenJSON(jsonObj) WITH (
        id          varchar(100) '$.id',
        source      varchar(100) '$.attributes.source',
        medium      varchar(100) '$.attributes.medium',
        visit_route nvarchar(MAX) '$.attributes.visit_route' AS JSON,
        keyword     varchar(100) '$.attributes.keyword'
    ) B

Returns

enter image description here

Upvotes: 2

GMB
GMB

Reputation: 222382

You would need to openjson() twice: once to unnest jsonObj, and another time to access the nested attributes:

select x.id, y.source, y.medium, y.visit_route
from [StageDB].[dbo].[LeadFeeder_visits_json] l
cross apply openjson(l.jsonObj) with(   
    id          nvarchar(100),
    attributes  nvarchar(max) as json
) x
cross apply openjson(x.attributes) with (
    source      nvarchar(100),
    medium      nvarchar(100),
    visit_route nvarchar(max) as json
) y

Note that you need as json when extracting nested json content.

Upvotes: 3

Related Questions