Parse JSON object in SQL with list

Question

I have a table with data structured like this. Each product ID has a list of element IDs--for each element, there is a dictionary including a list of elements and their assigned IDs. Not every element will have an ID on every product

product_id	element_id
product_1	{"FIRE": ["1630808"],"WATER": ["188028","234"],"SHADOW": ["213181"]

For each product I'd like to be able to count how many of each element ID appear, in a table like this:

product_id	fire_count	water_count	shadow_count	forest_count
product_1	1	2	1	0

I've tried using the LATERAL FLATTEN function with KEY and VALUE, but I'm getting duplicate results and wonder if there is a more crisp way of writing this type of query, especially because I also need to count instances where an ID does not appear.

My data is stored in Snowflake and I query it using Snowflake SQL.

Any advice? Thank you!

Please read : Why should I provide a Minimal Reproducible Example, even for a very simple SQL query? — MatBailie
– MatBailie, Commented Nov 25 at 5:50

Lukasz Szozda · Accepted Answer · 2025-11-25 15:07:47Z

2

It can be achieved without flattening and aggregation:

CREATE OR REPLACE TABLE TAB(PRODUCT_ID, ELEMENT_ID) AS SELECT 'product_1',
{'FIRE':['1630808'],'WATER':['188028','234'],'SHADOW':['213181']};

SELECT
  PRODUCT_ID,
  ARRAY_SIZE(ELEMENT_ID:FIRE) AS FIRE_COUNT,
  ARRAY_SIZE(ELEMENT_ID:WATER) AS WATER_COUNT,
  ARRAY_SIZE(ELEMENT_ID:SHADOW) AS SHADOW_COUNT,
FROM TAB;
/*
+------------+------------+-------------+--------------+
| PRODUCT_ID | FIRE_COUNT | WATER_COUNT | SHADOW_COUNT |
+------------+------------+-------------+--------------+
| product_1  |          1 |           2 |            1 |
+------------+------------+-------------+--------------+
*/

answered Nov 25 at 15:07

Lukasz Szozda

181k26 gold badges278 silver badges326 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

MatBailie Nov 25 at 15:45

I'd also suggest wrapping each value with COALESCE(..., 0)

MatBailie · Accepted Answer · 2025-11-25 05:46:31Z

1

I'd do the following...

flatten just one level
use ARRAY_SIZE() to count the elements in the arrays
use conditional aggregation to pivot the results

SELECT
  src.product_id,
  SUM(CASE WHEN f.key = 'fire'   THEN ARRAY_SIZE(f.value) ELSE 0 END)   AS fire_count,
  SUM(CASE WHEN f.key = 'water'  THEN ARRAY_SIZE(f.value) ELSE 0 END)   AS water_count,
  SUM(CASE WHEN f.key = 'shadow' THEN ARRAY_SIZE(f.value) ELSE 0 END)   AS shadow_count
FROM
  your_table   AS src
CROSS JOIN LATERAL
  FLATTEN (
    INPUT => src.element_id,
    OUTER => TRUE 
  )
    AS f
GROUP BY
  src.product_id

answered Nov 25 at 5:46

MatBailie

87.5k19 gold badges112 silver badges144 bronze badges

Collectives™ on Stack Overflow

Parse JSON object in SQL with list

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related