如何将 JSON 字典列表转换为 Snowflake 中的字符串列表？

Question

I have a Snowflake table which contains an id column and a favorite_books column. The favorite_books column contains a list of dictionaries in JSON. Each dictionary has a title and a ISBN attribute:

ID | favorite_books
--------------------
42 | [{"title": "LotR", "ISBN": "9780261102354"}, {"title": "HP1", "ISBN": "9780545582889"}]
13 | [{"title": "Faust", "ISBN": "9780192835956"}]

My real case is more complicated, but the part that is missing is to select only the titles. So the result should be:

ID | favorite_books
--------------------
42 | ["LotR", "HP1"]
13 | ["Faust"]

I saw Querying Semi-Structured Data , but I'm still not sure how to do that.

Answer 1

Adapting an existing answer for a similar data structure , you can try it this way (using a CTE, exploding the array, then grouping it back into a new array based on the ID):

with elements as (
    select ID, elements.value:title as title
    from table_name, LATERAL FLATTEN(input => favorite_books) elements
)
select elements.ID, ARRAY_AGG(elements.title) as favorite_books from elements
group by ID;

Or via a javascript UDF that's more straight-forward (just transforms the array):

create or replace function extract_titles(A array)
  returns array
  language javascript
  strict
  as
  $$
    return A.map(function(d) {return d.title});
  $$
;

select ID, EXTRACT_TITLES(favorite_books) as favorite_books from table_name;