Big Data

Join, share, and question the place your knowledge sits utilizing Amazon SageMaker Unified Studio

março 23, 2025
techhdesign


The flexibility for organizations to shortly analyze knowledge throughout a number of sources is essential for sustaining a aggressive benefit. Think about a state of affairs the place the retail analytics crew is making an attempt to reply a easy query: Amongst clients who bought summer season jackets final season, which clients are prone to have an interest within the new spring assortment?

Whereas the query is easy, getting the reply requires piecing collectively knowledge throughout a number of knowledge sources equivalent to buyer profiles saved in Amazon Easy Storage Service (Amazon S3) from buyer relationship administration (CRM) methods, historic buy transactions in an Amazon Redshift knowledge warehouse, and present product catalog info in Amazon DynamoDB. Historically, answering this query would contain a number of knowledge exports, advanced extract, remodel, and cargo (ETL) processes, and cautious knowledge synchronization throughout methods.

On this weblog put up, we’ll show how enterprise items can use Amazon SageMaker Unified Studio to find, subscribe to, and analyze these distributed knowledge property. By means of this unified question functionality, you’ll be able to create complete insights into buyer transaction patterns and buy habits for lively merchandise with out the standard obstacles of knowledge silos or the necessity to copy knowledge between methods.

SageMaker Unified Studio supplies a unified expertise for utilizing knowledge, analytics, and AI capabilities. You should utilize acquainted AWS providers for mannequin growth, generative AI, knowledge processing, and analytics—all inside a single, ruled surroundings. To strike a fantastic stability of democratizing knowledge and AI entry whereas sustaining strict compliance and regulatory requirements, Amazon SageMaker Information and AI Governance is constructed into SageMaker Unified Studio. With Amazon SageMaker Catalog, groups can collaborate by way of tasks, uncover, and entry accepted knowledge and fashions utilizing semantic search with generative AI-created metadata, or you need to use pure language to ask Amazon Q to seek out your knowledge. Inside SageMaker Unified Studio, organizations can implement a single, centralized permission mannequin with fine-grained entry controls, facilitating seamless knowledge and AI asset sharing by way of streamlined publishing and subscription workflows. Groups can even question the information instantly from sources equivalent to Amazon S3 and Amazon Redshift, by way of Amazon SageMaker Lakehouse.

SageMaker Lakehouse streamlines connecting to, cataloging, and managing permissions on knowledge from a number of sources. Constructed on AWS Glue Information Catalog and AWS Lake Formation, it organizes knowledge by way of catalogs that may be accessed by way of an open, Apache Iceberg REST API to assist guarantee safe entry to knowledge with constant, fine-grained entry controls. SageMaker Lakehouse organizes knowledge entry by way of two forms of catalogs: federated catalogs and managed catalogs (proven within the following determine). A catalog is a logical container that organizes objects from an information retailer, equivalent to schemas, tables, views, or materialized views equivalent to from Amazon Redshift. You may also create nested catalogs to reflect the hierarchical construction of your knowledge sources inside SageMaker Lakehouse.

  • Federated catalogs: By means of SageMaker Unified Studio, you’ll be able to create connections to exterior knowledge sources equivalent to Amazon DynamoDB. See Information connections in Amazon SageMaker Lakehouse for all of the supported exterior knowledge sources. These connections are saved within the AWS Glue Information Catalog (Information Catalog) and registered with Lake Formation, permitting you to create a federated catalog for every obtainable knowledge supply.
  • Managed catalogs: A managed catalog refers back to the knowledge that resides on Amazon S3 or Redshift Managed Storage (RMS).

The prevailing Information Catalog turns into the Default catalog (recognized by the AWS account quantity) and is available in SageMaker Lakehouse.

If the enterprise items don’t have an information warehouse however want the advantages of 1—equivalent to a question end result cache and question rewrite optimizations—then, they will create an RMS managed catalog in SageMaker Unified Studio. This can be a SageMaker Lakehouse managed catalog backed by RMS storage. The desk metadata is managed by Information Catalog. If you create an RMS managed catalog, it deploys an Amazon Redshift managed serverless workgroup. Customers can write knowledge to managed RMS tables utilizing Iceberg APIs, Amazon Redshift, or Zero-ETL ingestion from supported knowledge sources.

Useful working mannequin

In SageMaker Unified Studio, the infrastructure crew will allow the blueprints and configure the challenge profiles for instruments and applied sciences to the respective enterprise items to construct and monitor their pipelines. They may also onboard the groups to SageMaker Unified Studio, enabling them to construct the information merchandise in a single built-in, ruled surroundings. To implement standardization throughout the group, the central governance crew can even create hierarchical representations of enterprise items by way of area items and dictate sure actions that these groups can carry out underneath a website unit. International insurance policies equivalent to knowledge dictionaries (enterprise glossaries), knowledge classification tags, and extra info with metadata varieties might be created by the governance crew to make sure standardization and consistency throughout the group.

Particular person enterprise items will use these challenge profiles primarily based on their must course of the information utilizing the approved instrument of their selection and create knowledge merchandise. Enterprise items can benefit from the full flexibility to course of and devour the information with out worrying in regards to the upkeep of the underlying infrastructure. Relying on the character of the workloads, enterprise items can select a storage answer that most closely fits their use case. You should utilize SageMaker Lakehouse to unify the information throughout completely different knowledge sources.

To share the information outdoors the enterprise unit, the groups will publish the metadata of their knowledge to a SageMaker catalog and make it discoverable and accessible to different enterprise items. Amazon SageMaker Catalog serves as a central repository hub to retailer each technical and enterprise catalog info of the information product. To determine belief between the information producers and knowledge shoppers, SageMaker Catalog additionally integrates the knowledge high quality metrics and knowledge lineage occasions to trace and drive transparency in knowledge pipelines. Whereas sharing the information, knowledge producers of those enterprise items can apply fantastic grained entry management permissions at row and column stage to those property throughout subscription approval workflows. SageMaker Unified Studio mechanically grants subscription entry to the subscribed knowledge property after the subscription request is accepted by the information producer. As proven within the following determine, the information sharing functionality highlights that the information stays at its origin with the information producer, whereas shoppers from different enterprise items can devour and analyze it utilizing their very own compute assets. This strategy eliminates any knowledge duplication or knowledge motion.

Resolution overview

On this put up, we discover two eventualities for sharing knowledge between completely different groups (retail, advertising, and knowledge analysts). The answer on this put up provides you the implementation for a single account use case.

State of affairs 1

The retail crew must create a complete view of buyer habits to optimize their spring assortment launch. Their knowledge panorama is numerous:

  • Buyer profiles saved in Amazon S3 (default Information Catalog)
  • Historic buy transactions saved in RMS (SageMaker Lakehouse managed RMS catalog)
  • Stock info of the product in DynamoDB. (federated catalog)

The crew must share this unified view with their regional knowledge analysts whereas sustaining strict knowledge governance protocols. Information analysts uncover the information and subscribe to the information. We may also stroll by way of the publishing and subscription workflow as a part of the information sharing course of. To get a unified view of the shopper gross sales transactions for lively merchandise, the information analysts will use Amazon Athena.

Listed here are the excessive stage steps of the answer implementation as proven within the previous diagram:

  1. On this put up, we take an instance of two groups who take part within the collaboration. The retail crew has created a challenge retailsales-sql-project and the information analysts crew has created a challenge dataanalyst-sql-project inside SageMaker Unified Studio.
  2. The retail crew creates and shops their knowledge in numerous sources:
    1. buyer knowledge in Amazon S3 (comprises buyer knowledge)
    2. stock knowledge in a DynamoDB desk (comprises product catalog info)
    3. store_sales_lakehouse in SageMaker Lakehouse managed RMS (comprises buy historical past)
  3. The retail crew publishes the property to the challenge catalog to make them discoverable to different area members throughout the group.
  4. The information analysts crew discovers the information and subscribes to the information property.
  5. An incoming request is shipped to the retail crew, who then approves the subscription request. After the subscription is accepted, knowledge analysts use Athena to create a unified question from all of the subscribed knowledge property to get insights into the information.

On this state of affairs, we’ll evaluation how SageMaker Catalog manages the subscription grants to Information Catalog property (each federated and managed).

For this state of affairs, we assume that the retail crew doesn’t have their very own knowledge warehouse they usually wish to create and handle Amazon Redshift tables utilizing Information Catalog.

State of affairs 2

The advertising crew wants entry to transaction knowledge for marketing campaign optimization. They’ve marketing campaign efficiency knowledge saved in an Amazon Redshift knowledge warehouse. Nevertheless, to have improved marketing campaign ROI and higher useful resource allocation, they want knowledge from the retail crew to grasp precise buyer buy habits. To enhance the marketing campaign ROI, they want solutions to essential questions equivalent to:

  • What’s the true conversion fee throughout completely different buyer segments?
  • Which clients needs to be focused for upcoming promotions?
  • How do seasonal shopping for patterns have an effect on marketing campaign success?

Right here the retail crew shares the acquisition historical past knowledge store_sales to the advertising crew. On this state of affairs, proven within the previous determine, we assume that the retail crew has their very own knowledge warehouse and makes use of Amazon Redshift to retailer the acquisition historical past knowledge.

The excessive stage steps of the answer implementation for this state of affairs are:

  1. The advertising crew has created the challenge marketing-sql-project inside SageMaker Unified Studio.
  2. The retail crew has store_sales in Amazon Redshift knowledge warehouse (comprises buy historical past)
  3. The retail crew has revealed the property to the challenge catalog
  4. The advertising crew discovers the information and subscribes to the information property.
  5. An incoming request is shipped to the retail crew, who then approves the subscription request. After the subscription is accepted, the advertising crew makes use of Amazon Redshift to devour the acquisition historical past and determine high-value buyer segments.

On this state of affairs, we’ll evaluation the method of how SageMaker Catalog grants entry to managed Amazon Redshift property.

Stipulations

To comply with the step-by-step information, you should full the next conditions:

Be aware that the default SQL analytics challenge profile supplies you with a RedshiftServerless blueprint. Nevertheless, on this put up, we wish to showcase the information sharing capabilities of various kinds of SageMaker Lakehouse catalogs (managed and federated).

For the simplicity, we selected the SQL analytics challenge profile. Nevertheless, you may as well take a look at this through the use of the Customized challenge profile by deciding on particular blueprints equivalent to LakehouseCatalog and LakeHouseDatabase for eventualities the place the enterprise unit doesn’t have their very own knowledge warehouse.

Resolution walkthrough (State of affairs 1)

Step one focuses on making ready the information for every knowledge supply for unified entry.

Information preparation

On this part, you’ll create the next knowledge units:

  • buyer knowledge in Amazon S3 (default Information Catalog)
  • stock knowledge in a DynamoDB desk (federated catalog)
  • store_sales_lakehouse in SageMaker Lakehouse managed RMS (managed catalog)
  1. Check in to SageMaker Unified Studio as a member of the retail crew and choose the challenge retailsales-sql-project.
  2. On the highest menu, select Construct, and underneath DATA ANALYSIS & INTEGRATION, choose Question Editor.

  1. Choose the next choices:
    1. Beneath CONNECTIONS, choose Athena (Lakehouse).
    2. Beneath CATALOGS, choose AwsDataCatalog.
    3. Beneath DATABASES, choose glue_db_ or the shopper glue database title you offered throughout challenge creation.
    4. After the choices are chosen, select Select.

When customers choose a challenge profile inside SageMaker Unified Studio, the system mechanically triggers the related AWS CloudFormation stack (DataZone-Env-) and deploys the required infrastructure assets within the type of environments. Environments are the precise knowledge infrastructure behind a challenge.

  1. Run the next SQL:
CREATE TABLE buyer AS
SELECT 13251813 cust_id,'Joyce Deaton'   cust_name,'Greece'   cust_country, 'Joyce.Deaton@qhtrwert.edu'   cust_email
UNION
SELECT 1581546  ,'Daniel Dow'  ,'India'  , 'Daniel.Cass@hz05IuguG5b.org'  
UNION
SELECT 1581536  ,'Marie Lange'  ,'Canada'  , 'Marie.Lange@ka94on0lHy.edu'  
UNION
SELECT 1827661  ,'Wesley Harris'  ,'Rome'  , 'Wesley.Harris@c7NpgG4gyh.edu'  
UNION
SELECT 1581536  ,'Alexander Salyer'  ,'Germany'  , 'Alexander.Salyer@GxfK3iXetN.edu'  
UNION
SELECT 3581536  ,'Jerry Tracy'  ,'Swiss'  , 'Jerry.Tracy@VTtQp8OsUkv2hsygIh.edu'

  1. After the SQL is executed, you can see that the buyer desk has been created within the Lakehouse part underneath Lakehouse/AwsDataCatalog/glue_db_.

  1. The product catalog is saved in DynamoDB. You may create a brand new desk named stock in DynamoDB with partition key prod_id by way of AWS CloudShell with the next command:
aws dynamodb create-table 
    --table-name stock
    --attribute-definitions 
AttributeName=prod_id,AttributeType=N 
    --key-schema 
AttributeName=prod_id,KeyType=HASH 
    --provisioned-throughput 
ReadCapacityUnits=5,WriteCapacityUnits=5 
    --table-class STANDARD

  1. Populate the DynamoDB desk utilizing the next instructions:
aws dynamodb put-item --table-name stock --item '{"prod_id": {"N": "1"}, "prod_name": {"S": "Widget A"},"lively": {"S": "Y"}}' 

aws dynamodb put-item --table-name stock --item '{"prod_id": {"N": "2"}, "prod_name": {"S": "Gadget B"},"lively": {"S": "Y"}}'

aws dynamodb put-item --table-name stock --item '{"prod_id": {"N": "3"}, "prod_name": {"S": "Merchandise C"},"lively": {"S": "N"}}'

  1. To make use of the DynamoDB desk in SageMaker Unified Studio, you must configure a resource-based coverage that permits the suitable actions for the challenge function.
    1. To create the resource-based coverage, navigate to the DynamoDB console and select Tables from the navigation pane.
    2. Choose the Permissions desk and select Create desk coverage.

  1. The next is an instance coverage that permits connecting to DynamoDB tables as a federated supply. Change the  with the Area you’re engaged on,  with the AWS Account ID the place DynamoDB is deployed,  with the DynamoDB desk (on this case stock) that you simply intend to question from Amazon SageMaker Unified Studio and  with the Mission function Amazon Useful resource Title (ARN) in SageMaker Unified Studio portal. You will get the challenge function ARN by navigating to the challenge in SageMaker Unified Studio after which to Mission overview.

{
    "Model": "2012-10-17",
    "Assertion": [
        {
            "Effect": "Allow",
            "Principal": "*",
            "Action": [
                "dynamodb:Query",
                "dynamodb:Scan",
                "dynamodb:DescribeTable",
                "dynamodb:PartiQLSelect",
                "dynamodb:BatchWriteItem"
            ],
            "Useful resource": "arn:aws:dynamodb:::desk/",
            "Situation": {
                "ArnEquals": {
                    "aws:PrincipalArn": "arn:aws:iam:::function/"
                }
            }
        }
    ]
}

After the insurance policies are integrated on the DynamoDB desk, create an SageMaker Lakehouse connection inside SageMaker Unified Studio. As proven within the instance, dynamodb-connection-catalogs is created.

  1. After the connection is efficiently established, you will note the DynamoDB desk stock underneath Lakehouse.

The subsequent step is to create a managed catalog for RMS objects utilizing SageMaker Lakehouse.

  1. Select Information within the navigation pane.
  2. Within the knowledge explorer, select the plus icon so as to add an information supply.
  3. Choose Create Lakehouse catalog.
  4. Select Subsequent.

  1. Enter the title of the catalog. The catalog title offered within the instance is redshift-lakehouse-connection-catalogs. Select Add knowledge.

  1. After the connection is created, you will note the catalog underneath Lakehouse.

  1. This creates a managed Amazon Redshift Serverless workgroup in your AWS account. You will notice a brand new database dev@ within the managed Amazon Redshift Serverless workgroup.
    1. On the highest menu, select Construct, and underneath DATA ANALYSIS & INTEGRATION, choose Question Editor.
    2. Choose Redshift (Lakehouse) from CONNECTIONS, dev@ from DATABASES and public from SCHEMAS

  1. Run the next SQL so as. The SQL creates the store_sales_lakehouse desk within the dev database within the public schema. The retail crew inserts knowledge into the store_sales_lakehouse desk.
CREATE TABLE public.store_sales_lakehouse (
    sale_id INTEGER IDENTITY(1,1) PRIMARY KEY,
    cust_id INTEGER NOT NULL,
    sale_date DATE NOT NULL,
    sale_amount DECIMAL(10, 2) NOT NULL,
    prod_id INTEGER  NOT NULL,
    last_purchase_date DATE
);


INSERT INTO public.store_sales_lakehouse (cust_id, sale_date, sale_amount, prod_id, last_purchase_date)
VALUES
(13251813, '2023-01-15', 150.00, 1, '2023-01-15'),
(29033279, '2023-01-20', 200.00, 4, '2023-01-20'),
(12755125, '2023-02-01', 75.50, 3, '2023-02-01'),
(26009249, '2023-02-10', 300.00, 2, '2023-02-10'),
(3270685, '2023-02-15', 125.00, 2, '2023-02-15'),
(6520539, '2023-03-01', 100.00, 2, '2023-03-01'),
(10251183, '2023-03-10', 250.00, 1, '2023-03-10'),
(10251283, '2023-03-15', 180.00, 1, '2023-03-15'),
(10251383, '2023-04-01', 90.00, 2, '2023-04-01'),
(10251483, '2023-04-10', 220.00, 3, '2023-04-10'),
(10251583, '2023-04-15', 175.00, 3, '2023-04-15'),
(10251683, '2023-05-01', 130.00, 1, '2023-05-01'),
(10251783, '2023-05-10', 280.00, 1, '2023-05-10'),
(10251883, '2023-05-15', 195.00, 4, '2023-05-15'),
(10251983, '2023-06-01', 110.00, 2, '2023-06-01'),
(10251083, '2023-06-10', 270.00, 1, '2023-06-10'),
(10252783, '2023-06-15', 185.00, 2, '2023-06-15'),
(10253783, '2023-07-01', 95.00, 3, '2023-07-01'),
(10254783, '2023-07-10', 240.00, 1, '2023-07-10'),
(10255783, '2023-07-15', 160.00, 3, '2023-07-15');

  1. On profitable creation of the desk, you need to now be capable of question the information. Choose the desk store_sales_lakehouse and choose Question with Redshift.

Import property to the challenge catalog from numerous knowledge sources

To share your property outdoors your individual challenge to different enterprise items, you should first carry your metadata to SageMaker Catalog. To import the property into the challenge’s stock, you must create an information supply within the challenge catalog. On this part, we present you learn how to import the technical metadata from AWS Glue knowledge catalogs. Right here, you’ll import knowledge property from numerous sources that you’ve got created as a part of your knowledge preparation.

  1. Check in to SageMaker Unified Studio as a member of the retail crew. Choose the challenge retailsales-sql-project, underneath Mission catalog. Select Information sources and import the property by selecting Run.

  1. To import the federated catalog, create a brand new knowledge supply and select Run. This can import the metadata of the stock knowledge from DynamoDB desk.

  1. After profitable run of all the information sources, select Belongings underneath Mission catalog within the navigation airplane. You can find all of the property within the Stock of Mission catalog.

Publish the property

To make the property discoverable to the information analysts crew, the retail crew should publish their property.

  1. Within the challenge retailsales-sql-project, select Mission catalog and choose Belongings.
  2. Choose every asset within the INVENTORY tab, enrich the asset with the automated metadata era and PUBLISH ASSET.

Uncover the property

SageMaker Catalog inside SageMaker Unified Studio allows environment friendly knowledge asset discovery and entry administration. The information analysts crew indicators in to SageMaker Unified Studio and selects the challenge dataanalyst-sql-project. The information analysts crew then locates the specified property in SageMaker Catalog and initiates the subscription request.

On this part, members of dataanalyst-sql-project browse the catalog and discover the property. There are a number of methods to seek out the specified property.

  • Check in to SageMaker Unified Studio as a member of the information analysts crew. Select Uncover within the prime navigation bar and choose Catalog. Discover the specified asset by looking or coming into the title of the asset into the search bar.
  • Seek for the asset by way of a conversational interface utilizing Amazon Q.
  • Use the faceted filter search by deciding on the specified challenge within the BROWSE CATALOG.

The information analysts crew selects the challenge retailsales-sql-project.

Subscribe to the property

The information analysts crew submits a subscription request with an applicable justification for every of those property.

  1. For every asset, select SUBSCRIBE.
  2. Choose dataanalyst-sql-project in Mission.
  3. Present the Purpose for request as “want this knowledge for evaluation”.

Be aware that in the course of the subscription course of, the requester sees a message that the asset entry management and achievement will probably be Managed. Because of this SageMaker Unified Studio mechanically manages subscription entry grants and permissions for these property.

Subscription approval workflow

To approve the subscription request, you should be a member of the retail crew and choose the challenge that has revealed the asset.

  1. Check in to SageMaker Unified Studio as a member of the retail crew and choose the challenge retailsales-sql-project.
  2. Within the navigation pane, select Mission catalog after which choose Subscription requests.
  3. In INCOMING REQUESTS, select the REQUESTED tab and choose View request for every asset to see detailed info of the subscription request.

  • REQUEST DETAILS supplies details about the subscribing challenge, the requestor, and the justification to entry the asset.
  • RESPONSE DETAILS supplies an choice to approve the subscription with full entry to the information (Full entry) or restricted entry to the information (Approve with row or column filters). With restricted entry to knowledge, the subscription approval workflow course of provides granular entry management for delicate knowledge by way of row-level filtering and column-level filtering. Utilizing row filters, approvers can prohibit entry to particular information primarily based on outlined standards. Utilizing column filters, approvers can management entry to particular columns throughout the knowledge units. This permits excluding delicate fields whereas sharing the related knowledge. Approvers can implement these filters in the course of the approval course of, serving to to make sure that the information entry aligns with the group’s safety necessities and compliance insurance policies. For this put up, choose Full entry within the RESPONSE DETAILS
  • (Non-compulsory) Choice remark is the place you’ll be able to add a remark about accepting or rejecting the subscription request.
  • Select APPROVE.

  1. Repeat the subscription approval workflow course of for all of the requested property.
  2. After all of the subscription requests are accepted, select the APPROVED tab to view all of the accepted property.

Subscription achievement strategies

After subscription approval, a achievement course of manages entry to the property. SageMaker Unified Studio supplies achievement strategies for managed property and unmanaged property.

  • Managed property: SageMaker Unified Studio mechanically manages the achievement and permissions for property equivalent to AWS Glue tables and Amazon Redshift tables and views.
  • Unmanaged property: For unmanaged property, permissions are dealt with externally. SageMaker Unified Studio publishes normal occasions for actions equivalent to approvals by way of Amazon EventBridge, enabling integration with different AWS providers or third-party options for customized integrations.

On this state of affairs 1, as a result of the property are Information Catalogs, SageMaker Unified Studio grants and manages entry to those managed property in your behalf by way of Lake Formation. See the SageMaker Unified Studio subscription workflow for updates on sharing choices.

Analyze the information

The information analysts crew makes use of the subscribed knowledge property from various sources to get unified insights.

  1. As an information analyst, check in to SageMaker Unified Studio and choose the challenge dataanalyst-sql-project. Within the navigation pane, select Mission catalog and choose Belongings.
  2. Select the SUBSCRIBED tab to seek out all of the subscribed property from the retailsales-sql-project.
  3. The standing underneath every asset is Asset accessible. This means that the subscription grants are fulfilled and the information analysts crew can now devour the property with the compute of their selection.

Question utilizing Athena (subscription grants fulfilled utilizing Lake Formation)

As a member of the information analysts crew, create a unified view to get buy historical past with buyer info for lively merchandise.

  1. Within the dataanalyst-sql-project challenge, go to Construct and choose Question Editor.
  2. Use the next pattern question to get the required info. Change glue_db_ together with your subscribed glue database.
choose * from "redshift-lakehouse-connection-catalogs/dev"."public"."store_sales_lakehouse" gross sales 
 left  be part of "awsdatacatalog"."glue_db_"."buyer" buyer
 on gross sales.cust_id=buyer.cust_id
 internal  be part of "dynamodb-connection-catalogs"."default"."stock" stock
 on gross sales.prod_id = stock.prod_id
 the place stock.lively="Y"

Resolution walk-through (State of affairs 2)

On this state of affairs, we assume that the retail crew shops the acquisition historical past knowledge of their Amazon Redshift knowledge warehouse. Since you’re utilizing the default SQL analytics challenge profile to create the challenge, you’ll use a Redshift Serverless compute (challenge.redshift). The acquisition historical past knowledge is shared with the advertising crew for enhanced marketing campaign efficiency.

  1. Check in to SageMaker Unified Studio as a member of the retail crew and choose the challenge retailsales-sql-project.
  2. On the highest menu, select Construct, and underneath DATA ANALYSIS & INTEGRATION, choose Question Editor
  3. Choose the next choices:
    • Beneath CONNECTIONS, choose Redshift(Lakehouse).
    • Beneath CATALOGS, choose dev.
    • Beneath DATABASES, choose public.
  4. Run the next SQL:
CREATE TABLE public.store_sales (
sale_id INTEGER IDENTITY(1,1) PRIMARY KEY,
cust_id INTEGER NOT NULL,
sale_date DATE NOT NULL,
sale_amount DECIMAL(10, 2) NOT NULL,
prod_id INTEGER  NOT NULL,
last_purchase_date DATE
);

INSERT INTO public.store_sales (cust_id, sale_date, sale_amount, prod_id, last_purchase_date)
VALUES
(13251813, '2023-01-15', 150.00, 1, '2023-01-15'),
(29033279, '2023-01-20', 200.00, 4, '2023-01-20'),
(12755125, '2023-02-01', 75.50, 3, '2023-02-01'),
(26009249, '2023-02-10', 300.00, 2, '2023-02-10'),
(3270685, '2023-02-15', 125.00, 2, '2023-02-15'),
(6520539, '2023-03-01', 100.00, 2, '2023-03-01'),
(10251183, '2023-03-10', 250.00, 1, '2023-03-10'),
(10251283, '2023-03-15', 180.00, 1, '2023-03-15'),
(10251383, '2023-04-01', 90.00, 2, '2023-04-01'),
(10251483, '2023-04-10', 220.00, 3, '2023-04-10'),
(10251583, '2023-04-15', 175.00, 3, '2023-04-15'),
(10251683, '2023-05-01', 130.00, 1, '2023-05-01'),
(10251783, '2023-05-10', 280.00, 1, '2023-05-10'),
(10251883, '2023-05-15', 195.00, 4, '2023-05-15'),
(10251983, '2023-06-01', 110.00, 2, '2023-06-01'),
(10251083, '2023-06-10', 270.00, 1, '2023-06-10'),
(10252783, '2023-06-15', 185.00, 2, '2023-06-15'),
(10253783, '2023-07-01', 95.00, 3, '2023-07-01'),
(10254783, '2023-07-10', 240.00, 1, '2023-07-10'),
(10255783, '2023-07-15', 160.00, 3, '2023-07-15');

5. On profitable execution of the question, you will note store_sales underneath Redshift within the navigation pane.

Import the asset to the challenge catalog stock

To share your property outdoors your individual challenge to different advertising enterprise items, you should first share your metadata to SageMaker Catalog. To import the property into the challenge’s stock, you must run the information supply within the challenge catalog.

Within the challenge retailsales-sql-project, underneath Mission catalog, choose Information sources and import the asset store-sales. Choose the highlighted knowledge supply and select Run as proven within the screenshot.

Publish the asset

To make the property discoverable to the advertising crew, the retail crew should publish their asset.

  1. Go to the navigation pane and select Mission catalog, after which choose Belongings.
  2. Choose store-sales within the INVENTORY tab, enrich the asset with the automated metadata era and PUBLISH ASSET as illustrated within the screenshot.

Uncover and subscribe the asset

The advertising crew discovers and subscribes to the store-sales asset.

  1. Check in to SageMaker Unified Studio as a member of the advertising crew and choose marketing-sql-project.
  2. Navigate to the Uncover menu within the prime navigation bar and select Catalog. Discover the specified asset by looking or coming into the title of the asset into the search bar.
  3. Choose the asset and select SUBSCRIBE.
  4. Enter a justification in Purpose for request and select REQUEST.

Subscription approval workflow

The retail crew will get an incoming request of their challenge to approve the subscription request.

  1. Check in to the SageMaker Unified Studio and choose the challenge retailsales-sql-project as a member of the retail crew. Beneath Mission catalog, choose Subscription requests.
  2. Within the INCOMING REQUESTS, underneath the REQUESTED tab, choose View request for store-sales.

  1. You will notice detailed info for the subscription request.
  2. Choose Full entry within the RESPONSE DETAILS and select APPROVE.

Analyze the information

Check in to SageMaker Unified Studio as a member of the advertising crew and choose marketing-sql-project.

  1. Within the Mission catalog, choose Belongings and select the SUBSCRIBED tab to seek out all of the subscribed property from the retailsales-sql-project.
  2. Discover the standing underneath the asset marked as Asset accessible. This means that the subscription grants are fulfilled and the advertising crew can now devour the asset with the compute of their selection.

Question utilizing Amazon Redshift (subscription grants fulfilled utilizing native Amazon Redshift knowledge sharing)

To question the shared knowledge with Amazon Redshift compute, choose Construct after which Question Editor. Choose the next choices

  1. Beneath CONNECTIONS, choose Redshift(Lakehouse).
  2. Beneath CATALOGS, choose dev.
  3. Beneath DATABASES, choose challenge.
choose * from "dev"."challenge"."store_sales" gross sales

When a subscription to an Amazon Redshift desk or view is accepted, SageMaker Unified Studio mechanically provides the subscribed asset to the patron’s Amazon Redshift Serverless workgroup for the challenge. Discover the subscribed asset is shared underneath the folder challenge. Within the Redshift navigation pane, you may as well see the datashare created between the supply and the goal cluster. On this case, as a result of the information is shared in the identical account however between completely different clusters, SageMaker Unified Studio creates a view within the goal database and permissions are granted on the view. See Grant entry to managed Amazon Redshift property in Amazon SageMaker Unified Studio for details about knowledge sharing choices inside Amazon Redshift.

Clear up

Be sure to take away the SageMaker Unified Studio assets to keep away from any sudden prices. Begin by deleting the connections, catalogs, underlying knowledge sources, tasks, databases, and area that you simply created for this put up. For added particulars, see the Amazon SageMaker Unified Studio Administrator Information.

Conclusion

On this put up, we explored two distinct approaches to knowledge sharing and analytics.

Enterprise items with out an current knowledge warehouse can use a SageMaker Lakehouse managed RMS catalog. Within the first state of affairs, we showcased subscription achievement of AWS Glue Information Catalogs utilizing AWS Lake Formation for federated and managed catalogs. The information analysts crew was capable of join and subscribe to the information shared by the retail crew that resided in Amazon S3, Amazon Redshift, and different knowledge sources equivalent to DynamoDB by way of SageMaker Lakehouse.

Within the second state of affairs, we demonstrated the native data-sharing capabilities of Amazon Redshift. On this state of affairs, we assume that the retail crew has gross sales transactions saved in an Amazon Redshift knowledge warehouse. Utilizing the information sharing function of Amazon Redshift, the asset was shared to the advertising crew utilizing Amazon SageMaker Unified Studio.

Each approaches allow unified querying throughout various knowledge sources with groups capable of effectively uncover, publish, and subscribe to knowledge property whereas sustaining strict entry controls by way of Amazon SageMaker Information and AI Governance. Subscription achievement is automated, lowering the executive overhead. Utilizing the query-in-place strategy eliminates knowledge redundancy and maintains knowledge consistency whereas permitting unified evaluation throughout knowledge sources by way of a single built-in expertise.

To be taught extra, see the Amazon SageMaker Unified Studio Administrator Information and the next assets:

In regards to the authors

Lakshmi Nair is a Senior Analytics Specialist Options Architect at AWS. She focuses on designing superior analytics methods throughout industries. She focuses on crafting cloud-based knowledge platforms, enabling real-time streaming, huge knowledge processing, and sturdy knowledge governance. She might be reached by way of LinkedIn

Ramkumar Nottath is a Principal Options Architect at AWS specializing in Analytics providers. He enjoys working with numerous clients to assist them construct scalable, dependable huge knowledge and analytics options. His pursuits prolong to varied applied sciences equivalent to analytics, knowledge warehousing, streaming, knowledge governance, and machine studying. He loves spending time along with his household and associates.