Skip to content

CDI를 사용하여 계정 데이터 동기화하기

CDI를 사용하여 Braze 계정 데이터를 동기화하는 방법을 알아보세요.

전제 조건

CDI를 사용하여 계정 데이터를 동기화하려면 먼저 계정 스키마를 구성해야 합니다.

동기화 작동 방식

  • 각 동기화는 UPDATED_AT 이 마지막 동기화 타임스탬프보다 늦은 행을 가져옵니다.
  • 통합의 데이터는 제공된 id 을 기반으로 계정을 생성하거나 업데이트합니다.
  • DELETED 이 true 인 경우 계정이 삭제됩니다.
  • 동기화는 데이터 포인트를 기록하지는 않지만, 동기화된 모든 데이터는 총 저장된 데이터로 측정되는 총 계정 사용량에 포함되므로 변경된 데이터로만 제한할 필요가 없습니다.
  • 계정 스키마에 없는 필드는 삭제되므로 새 필드를 동기화하기 전에 스키마를 업데이트하세요.

계정 데이터 동기화하기

데이터 웨어하우스 또는 파일 저장소를 통해 CDI를 사용하여 계정 데이터를 동기화할 수 있습니다.

데이터 소스를 데이터 웨어하우스와 통합하려면:

  1. Snowflake에서 소스 테이블을 만듭니다. 예제의 이름을 사용하거나 고유한 데이터베이스, 스키마 및 테이블 이름을 선택합니다. 테이블 대신 뷰 또는 구체화된 뷰를 사용할 수도 있습니다. 
    1
2
3
4
5
6
7
8
9
10
11
12
13

     CREATE DATABASE BRAZE_CLOUD_PRODUCTION;
 CREATE SCHEMA BRAZE_CLOUD_PRODUCTION.INGESTION;
 CREATE OR REPLACE TABLE BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC (
      UPDATED_AT TIMESTAMP_NTZ(9) NOT NULL DEFAULT SYSDATE(),
      --ID of the account to be created or updated
      ID VARCHAR(16777216) NOT NULL,
      --Name of the account to be created or updated
      NAME VARCHAR(16777216) NOT NULL,
      --Account fields and values that should be added or updated
      PAYLOAD VARCHAR(16777216) NOT NULL,
      --The account associated with this ID should be deleted
      DELETED BOOLEAN
 );
  2. Create a role, warehouse, and user, and grant permissions. If you already have credentials from another sync, you can reuse them—just make sure they have access to the accounts table. 
    1
2
3
4
5
6
7
8
9
10
11

     CREATE ROLE BRAZE_INGESTION_ROLE;

 GRANT USAGE ON DATABASE BRAZE_CLOUD_PRODUCTION TO ROLE BRAZE_INGESTION_ROLE;
 GRANT USAGE ON SCHEMA BRAZE_CLOUD_PRODUCTION.INGESTION TO ROLE BRAZE_INGESTION_ROLE;
 GRANT SELECT ON TABLE BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC TO ROLE BRAZE_INGESTION_ROLE;

 CREATE WAREHOUSE BRAZE_INGESTION_WAREHOUSE;
 GRANT USAGE ON WAREHOUSE BRAZE_INGESTION_WAREHOUSE TO ROLE BRAZE_INGESTION_ROLE;

 CREATE USER BRAZE_INGESTION_USER;
 GRANT ROLE BRAZE_INGESTION_ROLE TO USER BRAZE_INGESTION_USER;
  3. If you use network policies, allowlist the Braze IPs so the CDI service can connect. For the list of IPs, see Cloud Data Ingestion.
  4. In the Braze dashboard, go to Data Settings > Cloud Data Ingestion and create a new sync.
  5. Enter connection details (or reuse existing ones), then add the source table.
  6. Select the Accounts sync type, then enter the integration name and schedule.
  7. Choose the sync frequency.
  8. Add the public key from the dashboard to the user you created. This requires a user with SECURITYADMIN access or higher in Snowflake.
  9. Select Test Connection to confirm the setup.
  10. When you’re finished, save the sync.
  1. Create a source table in Redshift. Use the names in the example or choose your own database, schema, and table names. You can also use a view or materialized view instead of a table. 
    1
2
3
4
5
6
7
8
9
10
11
12
13

     CREATE DATABASE BRAZE_CLOUD_PRODUCTION;
 CREATE SCHEMA BRAZE_CLOUD_PRODUCTION.INGESTION;
 CREATE TABLE BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC (
    updated_at timestamptz default sysdate not null,
    --ID of the account to be created or updated
    id varchar not null,
    --Name of the account to be created or updated
    name varchar not null,
    --Account fields and values that should be added or updated
    payload varchar(max),
    --The account associated with this ID should be deleted
    deleted boolean
 )

  2. Create a user and grant permissions. If you already have credentials from another sync, you can reuse them—just make sure they have access to the accounts table.

    1
2
3

     CREATE USER braze_user PASSWORD '{password}';
 GRANT USAGE ON SCHEMA BRAZE_CLOUD_PRODUCTION.INGESTION to braze_user;
 GRANT SELECT ON TABLE ACCOUNTS_SYNC TO braze_user;
  3. If you have a firewall or network policies, allow Braze access to your Redshift instance. For the list of IPs, see Cloud Data Ingestion.
  1. (Optional) Create a new project or dataset for your source table. 
    1

     CREATE SCHEMA BRAZE-CLOUD-PRODUCTION.INGESTION;
  2. Create the source table for your CDI integration: 
    1
2
3
4
5
6
7
8

     CREATE TABLE `BRAZE-CLOUD-PRODUCTION.INGESTION.ACCOUNTS_SYNC`
 (
   updated_at TIMESTAMP DEFAULT current_timestamp,
   id STRING,
   name STRING,
   payload JSON,
   deleted BOOLEAN
 );

    Refer to the following when creating your source table:

  1. Create a user and grant permissions. If you already have credentials from another sync, you can reuse them as long as they have access to the accounts table.

    After granting permissions, generate a JSON key. See Keys create and delete for instructions. You’ll upload it in the Braze dashboard later.

  1. If you use network policies, allow Braze IPs to access your BigQuery instance. For the list of IPs, see Cloud Data Ingestion.
  1. Create a catalog or schema for your source table. 
    1

     CREATE SCHEMA BRAZE-CLOUD-PRODUCTION.INGESTION;
  2. Create the source table for your CDI integration: 
    1
2
3
4
5
6
7
8

     CREATE TABLE `BRAZE-CLOUD-PRODUCTION.INGESTION.ACCOUNTS_SYNC`
 (
   updated_at TIMESTAMP DEFAULT current_timestamp(),
   id STRING,
   name STRING,
   payload STRING, STRUCT, or MAP,
   deleted BOOLEAN
 );

    Refer to the following when creating your source table:

  1. Create a personal access token in Databricks:
    1. Select your username, then select User Settings.
    2. On the Access tokens tab, select Generate new token.
    3. Add a comment to identify the token, such as “Braze CDI”.
    4. Leave Lifetime (days) blank for no expiration, then select Generate.
    5. Copy and save the token securely for use in the Braze dashboard.
  1. If you use network policies, allow Braze IPs to access your Databricks instance. For the list of IPs, see Cloud Data Ingestion.
  1. Create one or more tables for your CDI integration with these fields: 
    1
2
3
4
5
6
7
8
9

     CREATE OR ALTER TABLE [warehouse].[schema].[CDI_table_name] 
 (
   UPDATED_AT DATETIME2(6) NOT NULL,
   PAYLOAD VARCHAR NOT NULL,
   ID VARCHAR NOT NULL,
   NAME VARCHAR NOT NULL,
   DELETED BIT
 )
 GO
  1. Create a service principal and grant permissions. If you already have credentials from another sync, you can reuse them—just make sure they have access to the accounts table.
  1. If you use network policies, allow Braze IPs to access your Microsoft Fabric instance. For the list of IPs, see Cloud Data Ingestion.

To sync account data from file storage, create a source file with the following fields.

The following examples show valid JSON and CSV formats for syncing account data from file storage.

1
2
3
4

{"id":"s3-qa-0","name":"account0","payload":"{\"attribute_0\": \"GT896\", \"attribute_1\": 74, \"attribute_2\": true, \"retention\": {\"previous_purchases\": 21, \"vip\": false}, \"last_visit\": \"2023-08-08T16:03:26.600803\"}"}
{"id":"s3-qa-1","name":"account1","payload":"{\"attribute_0\": \"GT896\", \"attribute_1\": 74, \"attribute_2\": true, \"retention\": {\"previous_purchases\": 21, \"vip\": false}, \"last_visit\": \"2023-08-08T16:03:26.600803\"}","deleted":true}
{"id":"s3-qa-2","name":"account2","payload":"{\"attribute_0\": \"GT896\", \"attribute_1\": 74, \"attribute_2\": true, \"retention\": {\"previous_purchases\": 21, \"vip\": false}, \"last_visit\": \"2023-08-08T16:03:26.600803\"}","deleted":false}
{"id":"s3-qa-3","name":"account3","payload":"{\"attribute_0\": \"GT896\", \"attribute_1\": 74, \"attribute_2\": true, \"retention\": {\"previous_purchases\": 21, \"vip\": false}, \"last_visit\": \"2023-08-08T16:03:26.600803\"}"}

1
2
3

ID,NAME,PAYLOAD,DELETED
85,"ACCOUNT_1","{""region"": ""APAC"", ""employees"": 850}",TRUE 
1,"ACCOUNT_2","{""region"": ""EMEA"", ""employees"": 10000}",FALSE

1
2
3

ID,NAME,PAYLOAD
85,"ACCOUNT_1","{""region"": ""APAC"", ""employees"": 850}"
1,"ACCOUNT_2","{""region"": ""EMEA"", ""employees"": 10000}"

동기화 보기 만들기

데이터 웨어하우스에서 동기화 보기를 만들면 추가 쿼리를 다시 작성할 필요 없이 소스를 자동으로 새로고침할 수 있습니다.

예를 들어 account_id, account_name, 그리고 3개의 추가 속성이 있는 account_details_1 라는 계정 데이터 테이블이 있는 경우 다음과 같은 동기화 보기를 만들 수 있습니다:

1
2
3
4
5
6
7
8
9
10
11
12
13
14

CREATE VIEW BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC AS 
SELECT
    CURRENT_TIMESTAMP as UPDATED_AT,
    account_id as id,
    account_name as name,
    TO_JSON(
        OBJECT_CONSTRUCT (
            'attribute_1',
            attribute_1,
            'attribute_2',
            attribute_2,
            'attribute_3',
            attribute_3)
    )as PAYLOAD FROM "account_details_1";

1
2
3
4
5
6
7
8
9
10
11
12
13
14

CREATE TABLE BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC AS
SELECT
    CURRENT_TIMESTAMP as UPDATED_AT,
    account_id as id,
    account_name as name,
    JSON_SERIALIZE(
        OBJECT (
            'attribute_1',
            attribute_1,
            'attribute_2',
            attribute_2,
            'attribute_3',
            attribute_3)
    ) as PAYLOAD FROM "account_details_1";

1
2
3
4
5
6
7
8
9
10
11
12

CREATE view IF NOT EXISTS BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC AS (SELECT
    last_updated as UPDATED_AT,
    account_id as ID,
    account_name as NAME,
    TO_JSON(
      STRUCT(
      attribute_1,
      attribute_2,
      attribute_3,
      )
    ) as PAYLOAD 
  FROM `BRAZE_CLOUD_PRODUCTION.INGESTION.account_details_1`);

1
2
3
4
5
6
7
8
9
10
11
12

CREATE view IF NOT EXISTS BRAZE_CLOUD_PRODUCTION.INGESTION.ACCOUNTS_SYNC AS (SELECT
    last_updated as UPDATED_AT,
    account_id as ID,
    account_name as NAME,
    TO_JSON(
      STRUCT(
      attribute_1,
      attribute_2,
      attribute_3,
      )
    ) as PAYLOAD 
  FROM `BRAZE_CLOUD_PRODUCTION.INGESTION.account_details_1`);

1
2
3
4
5
6
7
8

CREATE VIEW [BRAZE_CLOUD_PRODUCTION].[INGESTION].[ACCOUNTS_SYNC]
AS SELECT 
    account_id as ID,
    account_name as NAME,
    CURRENT_TIMESTAMP as UPDATED_AT,
    JSON_OBJECT('attribute_1':attribute_1, 'attribute_2':attribute_2, 'attribute_3':attribute_3, 'attribute_4':attribute_4) as PAYLOAD

FROM [braze].[account_details_1] ;
Github   GitHub 에서 이 페이지를 편집합니다.
New Stuff!