Utilisation du masquage dynamique des données avec des chemins de type de données SUPER

Les traductions sont fournies par des outils de traduction automatique. En cas de conflit entre le contenu d'une traduction et celui de la version originale en anglais, la version anglaise prévaudra.

Les traductions sont fournies par des outils de traduction automatique. En cas de conflit entre le contenu d'une traduction et celui de la version originale en anglais, la version anglaise prévaudra.

Amazon Redshift permet d’attacher des politiques de masquage dynamique des données aux chemins des colonnes de type SUPER. Pour plus d’informations sur le type de données SUPER, consultez Données semi-structurées dans Amazon Redshift.

Lorsque vous associez des politiques de masquage aux chemins de colonnes de type SUPER, tenez compte des points suivants.

Exemples

Association de politiques de masquage aux chemins SUPER

L’exemple suivant attache plusieurs politiques de masquage à plusieurs chemins de type SUPER dans une colonne.

CREATE TABLE employees (
    col_person SUPER
);

INSERT INTO employees
VALUES
    (
        json_parse('
            {
                "name": {
                    "first": "John",
                    "last": "Doe"
                },
                "age": 25,
                "ssn": "111-22-3333",
                "company": "Company Inc."
            }
        ')
    ),
    (
        json_parse('
            {
                "name": {
                    "first": "Jane",
                    "last": "Appleseed"
                },
                "age": 34,
                "ssn": "444-55-7777",
                "company": "Organization Org."
            }
        ')
    )
;
GRANT ALL ON ALL TABLES IN SCHEMA "public" TO PUBLIC;

-- Create the masking policies.

-- This policy converts the given name to all uppercase letters.
CREATE MASKING POLICY mask_first_name
WITH(first_name TEXT)
USING ( UPPER(first_name) );

-- This policy replaces the given name with the fixed string 'XXXX'.
CREATE MASKING POLICY mask_last_name
WITH(last_name TEXT)
USING ( 'XXXX'::TEXT );

-- This policy rounds down the given age to the nearest 10.
CREATE MASKING POLICY mask_age
WITH(age INT)
USING ( (FLOOR(age::FLOAT / 10) * 10)::INT );

-- This policy converts the first five digits of the given SSN to 'XXX-XX'.
CREATE MASKING POLICY mask_ssn
WITH(ssn TEXT)
USING ( 'XXX-XX-'::TEXT || SUBSTRING(ssn::TEXT FROM 8 FOR 4) );

-- Attach the masking policies to the employees table.
ATTACH MASKING POLICY mask_first_name
ON employees(col_person.name.first)
TO PUBLIC;

ATTACH MASKING POLICY mask_last_name
ON employees(col_person.name.last)
TO PUBLIC;

ATTACH MASKING POLICY mask_age
ON employees(col_person.age)
TO PUBLIC;

ATTACH MASKING POLICY mask_ssn
ON employees(col_person.ssn)
TO PUBLIC;

-- Verify that your masking policies are attached.
SELECT
    policy_name,
    TABLE_NAME,
    priority,
    input_columns,
    output_columns
FROM
    svv_attached_masking_policy;

   policy_name   | table_name | priority |           input_columns           |          output_columns
-----------------+------------+----------+-----------------------------------+-----------------------------------
 mask_age        | employees  |        0 | ["col_person.\"age\""]            | ["col_person.\"age\""]
 mask_first_name | employees  |        0 | ["col_person.\"name\".\"first\""] | ["col_person.\"name\".\"first\""]
 mask_last_name  | employees  |        0 | ["col_person.\"name\".\"last\""]  | ["col_person.\"name\".\"last\""]
 mask_ssn        | employees  |        0 | ["col_person.\"ssn\""]            | ["col_person.\"ssn\""]
(4 rows)

-- Observe the masking policies taking effect.
SELECT col_person FROM employees ORDER BY col_person.age;

-- This result is formatted for ease of reading.
         col_person
--------------------------------
{
    "name": {
        "first": "JOHN",
        "last": "XXXX"
    },
    "age": 20,
    "ssn": "XXX-XX-3333",
    "company": "Company Inc."
}
{
    "name": {
        "first": "JANE",
        "last": "XXXX"
    },
    "age": 30,
    "ssn": "XXX-XX-7777",
    "company": "Organization Org."
}

Voici quelques exemples d’attachements non valides de politique de masquage à des chemins SUPER.

-- This attachment fails because there is already a policy
-- with equal priority attached to employees.name.last, which is
-- on the same SUPER path as employees.name.
ATTACH MASKING POLICY mask_ssn
ON employees(col_person.name)
TO PUBLIC;
ERROR:  DDM policy "mask_last_name" is already attached on relation "employees" column "col_person."name"."last"" with same priority
               
-- Create a masking policy that masks DATETIME objects.
CREATE MASKING POLICY mask_date
WITH(INPUT DATETIME)
USING ( INPUT );
               
-- This attachment fails because SUPER type columns can't contain DATETIME objects.
ATTACH MASKING POLICY mask_date
ON employees(col_person.company)
TO PUBLIC;
ERROR:  cannot attach masking policy for output of type "timestamp without time zone" to column "col_person."company"" of type "super

Voici un exemple d’attachement d’une politique de masquage à un chemin SUPER qui n’existe pas. Par défaut, Amazon Redshift résoudra le chemin à la valeur NULL.

ATTACH MASKING POLICY mask_first_name
ON employees(col_person.not_exists)
TO PUBLIC;

SELECT col_person FROM employees LIMIT 1;

-- This result is formatted for ease of reading.
         col_person
-----------------------------------
{
    "name": {
        "first": "JOHN",
        "last": "XXXX"
    },
    "age": 20,
    "ssn": "XXX-XX-3333",
    "company": "Company Inc.",
    "not_exists": null
}