Return to Answer

• It is always better to present real code in your question - like actual CREATE TABLE statements, a syntactically correct JSON snippet and a query that actually can run.
• For translating JSON arrays into PostgreSQL arrays, I used here a technique that is described in detail in Erwin Brandstetter's excellent answer.
• Check the documentation about the different JSON(B) operators and functions. It contains quite a lot of useful stuff.

• It is always better to present real code in your question - like actual CREATE TABLE statements, a syntactically correct JSON snippet and a query that actually can run.
• For translating JSON arrays into PostgreSQL arrays, I used here a technique that is described in detail in Erwin Brandstetter's excellent answer.
• Check the documentation about the different JSON(B) operators and functions. It contains quite a lot of useful stuff.

Notes:

• It is always better to present real code in your question - like actual CREATE TABLE statements, a syntactically correct JSON snippet and a query that actually can run.
• For translating JSON arrays into PostgreSQL arrays, I used here a technique that is described in detail in Erwin Brandstetter's excellent answer.
• Check the documentation about the different JSON(B) operators and functions. It contains quite a lot of useful stuff.

WITH condition_elements (... like above ...)
SELECT data, key, value 
  FROM some_data, condition_elements
 WHERE data->key=ANY(value);

       data       │  key   │         value          
──────────────────┼────────┼────────────────────────
 "access"=>"user" │ access │ {superuser,user,login}

-- add some more data for testing
INSERT INTO some_data
VALUES ('departement=>bla,access=>bla'),
       ('departement=>it,access=>bla'),
       ('departement=>it,access=>login');

WITH condition_elements (... as above ...)
SELECT *
  FROM (
        SELECT data, bool_and(data->key=ANY(value)) AS full_match
          FROM some_data, condition_elements
         GROUP BY data
       ) t
 WHERE full_match;

                  data                  │ full_match 
────────────────────────────────────────┼────────────
 "access"=>"login", "departement"=>"it" │ t
 "access"=>"user"                       │ t