snowflake-cloud-data-platformsnowflake-schema

Dynamically lateral flatten in Snowflake


{
  "explosives_UG":{
   "isCritical": false,
   "value": "N/A"
  },
  "explosivesUG": {
   "comment": "Test 619 313",
   "createWo": false,
   "isCritical": false,
   "value": "No"
  },
  "generalAttachment": null,
  "generalComment": "Test 619 313",
  "guardsPostedUG": {
    "isCritical": false,
    "value": "Yes"
  },
  "imminentUG": {
    "isCritical": false,
    "value": "Yes"
  }
}

using the above JSON format, output is generated as below via Snowflake query

question        answer
explosives_UG   N/A
explosivesUG    No
guardsPostedUG  Yes
imminentUG      Yes
SELECT 
  kv.key AS question,
  kv.value:value AS answer
FROM 
  json_data,
  LATERAL FLATTEN(input => json_col) kv
WHERE 
  kv.key IN ('explosives_UG', 'explosivesUG', 'guardsPostedUG', 'imminentUG');

Question : We will be getting lot of questions/answers and handling them manually in the code is time consuming.. How to use dynamically extract JSON values using LATERAL FLATTEN, can somebody help please...


Solution

  • so if we slam that JSON into a single flatten:

    with cte_json_data as (
    select 
        PARSE_JSON('
        {
      "explosives_UG":{
       "isCritical": false,
       "value": "N/A"
      },
      "explosivesUG": {
       "comment": "Test 619 313",
       "createWo": false,
       "isCritical": false,
       "value": "No"
      },
      "generalAttachment": null,
      "generalComment": "Test 619 313",
      "guardsPostedUG": {
        "isCritical": false,
        "value": "Yes"
      },
      "imminentUG": {
        "isCritical": false,
        "value": "Yes"
      }
    }') as json
    )
    select
        m.key as questions
        ,m.value
    from cte_json_data as j
        ,table(flatten(input=>j.json)) as m
    

    we get:

    enter image description here

    and you want "it appears" the objects that have a known object, thus we can use IS_NULL_VALUE:

    with cte_json_data as (
    select 
        PARSE_JSON('
        {
      "explosives_UG":{
       "isCritical": false,
       "value": "N/A"
      },
      "explosivesUG": {
       "comment": "Test 619 313",
       "createWo": false,
       "isCritical": false,
       "value": "No"
      },
      "generalAttachment": null,
      "generalComment": "Test 619 313",
      "guardsPostedUG": {
        "isCritical": false,
        "value": "Yes"
      },
      "imminentUG": {
        "isCritical": false,
        "value": "Yes"
      }
    }') as json
    )
    select
        m.key as questions
        ,m.value:value as answer
    from cte_json_data as j
        ,table(flatten(input=>j.json)) as m
    where is_null_value(m.value:isCritical) = false;
    

    enter image description here

    would appear to do what you are after.