❄️ Snowflake Integration Guide

Connect your Snowflake account to mimilabs data using Delta Direct — Snowflake's native ability to read Delta tables from S3.

Enterprise Feature: Snowflake integration is available with an Enterprise license. Contact info@mimilabs.ai to get started.

Already using Databricks? Delta Sharing is even simpler — contact us for setup.

How It Works

Your Snowflake account assumes an IAM role we provision, giving it read-only access to the specific tables you've requested.

┌─────────────────┐                              ┌─────────────────┐
│   Your          │        assumes role          │   mimilabs      │
│   Snowflake     │  ──────────────────────────► │   AWS           │
│   Account       │                              │                 │
└─────────────────┘                              └────────┬────────┘
                                                          │ reads
                                                          ▼
                                                 ┌─────────────────┐
                                                 │   S3 (Delta)    │
                                                 │   mimilabs data │
                                                 └─────────────────┘

Step 1: Request Enterprise Access

Your organization and use case
Which tables or schemas you need

Once your Enterprise license is set up, we'll provision access and send you:

S3 path(s) for your requested tables
IAM Role ARN to assume

Request Enterprise Access

Step 2: Create Catalog Integration

Run this once in your Snowflake account (requires ACCOUNTADMIN role):

-- Run this once per Snowflake account
CREATE OR REPLACE CATALOG INTEGRATION mimi_ws_1
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = DELTA
  ENABLED = TRUE;

Step 3: Create External Volume

We'll provide you with the specific S3 path and IAM Role ARN. Use them in this template:

CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = '{{S3_PATH_PROVIDED_BY_MIMILABS}}'
      STORAGE_AWS_ROLE_ARN = '{{IAM_ROLE_ARN_PROVIDED_BY_MIMILABS}}'
    )
  )
  ALLOW_WRITES = FALSE;

Example with real values:

CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = 's3://mimilabs-data-share/client-abc/nppes/'
      STORAGE_AWS_ROLE_ARN = 'arn:aws:iam::533267369554:role/mimilabs-client-abc-role'
    )
  )
  ALLOW_WRITES = FALSE;

Step 4: Send Us Your Snowflake Credentials

After creating the external volume, run:

DESC EXTERNAL VOLUME mimilabs_volume;

Send us these two values from the output:

STORAGE_AWS_IAM_USER_ARN
STORAGE_AWS_EXTERNAL_ID

We'll update our IAM trust policy to authorize your Snowflake account.

Step 5: Create Database and Schema

Set up where your mimilabs tables will live:

CREATE DATABASE IF NOT EXISTS mimi_ws_1;
CREATE SCHEMA IF NOT EXISTS mimi_ws_1.nppes;  -- or other schema name
USE DATABASE mimi_ws_1;
USE SCHEMA nppes;

Tip: We recommend using mimi_ws_1 as your database name and matching schema names (e.g., nppes, fda, cmscoding) to stay consistent with our Databricks environment. This way, SQL queries generated by mimibot can be used directly in your Snowflake without modification.

Step 6: Create Iceberg Tables

Once we confirm the IAM trust policy is updated, create tables for each dataset:

Single Table Setup

CREATE ICEBERG TABLE {{TABLE_NAME}}
  BASE_LOCATION = ''
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

Multiple Tables (Shared Parent Path)

If we've set up your external volume to point to a parent directory containing multiple tables:

-- NPPES Provider Directory
CREATE ICEBERG TABLE npidata
  BASE_LOCATION = 'npidata/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- FDA NDC Directory
CREATE ICEBERG TABLE ndc_directory
  BASE_LOCATION = 'ndc_directory/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- HCPCS Codes
CREATE ICEBERG TABLE hcpcs
  BASE_LOCATION = 'hcpcs/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

Step 7: Query Your Data

-- Query NPPES provider directory
SELECT npi, provider_first_name, provider_last_name_legal_name,
       provider_business_practice_location_address_city_name,
       provider_business_practice_location_address_state_name
FROM mimi_ws_1.nppes.npidata
LIMIT 10;

Quick Reference: Complete Setup Script

Copy and customize this full script:

-- ============================================
-- MIMILABS SNOWFLAKE INTEGRATION SETUP
-- ============================================

-- 1. Create catalog integration (one-time)
CREATE OR REPLACE CATALOG INTEGRATION mimi_ws_1
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = DELTA
  ENABLED = TRUE;

-- 2. Create external volume
--    Replace placeholders with values from mimilabs
CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = '{{S3_PATH}}'
      STORAGE_AWS_ROLE_ARN = '{{IAM_ROLE_ARN}}'
    )
  )
  ALLOW_WRITES = FALSE;

-- 3. Get credentials to send to mimilabs
DESC EXTERNAL VOLUME mimilabs_volume;
-- >>> Send STORAGE_AWS_IAM_USER_ARN and STORAGE_AWS_EXTERNAL_ID to mimilabs <<<

-- 4. Create database/schema (use mimi_ws_1 for mimibot query compatibility)
CREATE DATABASE IF NOT EXISTS mimi_ws_1;
CREATE SCHEMA IF NOT EXISTS mimi_ws_1.{{SCHEMA_NAME}};

-- 5. Create tables (after mimilabs confirms IAM setup)
CREATE ICEBERG TABLE mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}}
  BASE_LOCATION = '{{BASE_LOCATION}}'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- 6. Query (same syntax as mimibot/Databricks)
SELECT * FROM mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}} LIMIT 10;

Refreshing Data

With AUTO_REFRESH = TRUE, your Iceberg tables automatically detect new data — no action required.

To manually trigger a refresh:

ALTER ICEBERG TABLE mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}} REFRESH;

To check the last refresh time:

SHOW ICEBERG TABLES LIKE '{{TABLE_NAME}}';

Troubleshooting

`not authorized to perform: sts:AssumeRole`

Cause: IAM trust policy not configured
Solution: Send us your STORAGE_AWS_IAM_USER_ARN and STORAGE_AWS_EXTERNAL_ID

`Failed to find the metadata of Delta table`

Cause: Wrong path or BASE_LOCATION
Solution: Verify BASE_LOCATION matches the table subdirectory

`Forbidden (Status Code: 403)`

Cause: IAM permissions issue
Solution: Contact mimilabs — we'll check the policy

`Invalid value for parameter 'CATALOG_SOURCE'`

Cause: Snowflake edition limitation
Solution: Delta Direct requires Snowflake Enterprise Edition or higher

Need Help?

Your Snowflake account identifier
Error messages (if any)
Which tables you're trying to access

❄️ Snowflake Integration Guide

Connect your Snowflake account to mimilabs data using Delta Direct — Snowflake's native ability to read Delta tables from S3.

Enterprise Feature: Snowflake integration is available with an Enterprise license. Contact info@mimilabs.ai to get started.

Already using Databricks? Delta Sharing is even simpler — contact us for setup.

How It Works

Your Snowflake account assumes an IAM role we provision, giving it read-only access to the specific tables you've requested.

┌─────────────────┐                              ┌─────────────────┐
│   Your          │        assumes role          │   mimilabs      │
│   Snowflake     │  ──────────────────────────► │   AWS           │
│   Account       │                              │                 │
└─────────────────┘                              └────────┬────────┘
                                                          │ reads
                                                          ▼
                                                 ┌─────────────────┐
                                                 │   S3 (Delta)    │
                                                 │   mimilabs data │
                                                 └─────────────────┘

Step 1: Request Enterprise Access

Your organization and use case
Which tables or schemas you need

Once your Enterprise license is set up, we'll provision access and send you:

S3 path(s) for your requested tables
IAM Role ARN to assume

Request Enterprise Access

Step 2: Create Catalog Integration

Run this once in your Snowflake account (requires ACCOUNTADMIN role):

-- Run this once per Snowflake account
CREATE OR REPLACE CATALOG INTEGRATION mimi_ws_1
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = DELTA
  ENABLED = TRUE;

Step 3: Create External Volume

We'll provide you with the specific S3 path and IAM Role ARN. Use them in this template:

CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = '{{S3_PATH_PROVIDED_BY_MIMILABS}}'
      STORAGE_AWS_ROLE_ARN = '{{IAM_ROLE_ARN_PROVIDED_BY_MIMILABS}}'
    )
  )
  ALLOW_WRITES = FALSE;

Example with real values:

CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = 's3://mimilabs-data-share/client-abc/nppes/'
      STORAGE_AWS_ROLE_ARN = 'arn:aws:iam::533267369554:role/mimilabs-client-abc-role'
    )
  )
  ALLOW_WRITES = FALSE;

Step 4: Send Us Your Snowflake Credentials

After creating the external volume, run:

DESC EXTERNAL VOLUME mimilabs_volume;

Send us these two values from the output:

STORAGE_AWS_IAM_USER_ARN
STORAGE_AWS_EXTERNAL_ID

We'll update our IAM trust policy to authorize your Snowflake account.

Step 5: Create Database and Schema

Set up where your mimilabs tables will live:

CREATE DATABASE IF NOT EXISTS mimi_ws_1;
CREATE SCHEMA IF NOT EXISTS mimi_ws_1.nppes;  -- or other schema name
USE DATABASE mimi_ws_1;
USE SCHEMA nppes;

Tip: We recommend using mimi_ws_1 as your database name and matching schema names (e.g., nppes, fda, cmscoding) to stay consistent with our Databricks environment. This way, SQL queries generated by mimibot can be used directly in your Snowflake without modification.

Step 6: Create Iceberg Tables

Once we confirm the IAM trust policy is updated, create tables for each dataset:

Single Table Setup

CREATE ICEBERG TABLE {{TABLE_NAME}}
  BASE_LOCATION = ''
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

Multiple Tables (Shared Parent Path)

If we've set up your external volume to point to a parent directory containing multiple tables:

-- NPPES Provider Directory
CREATE ICEBERG TABLE npidata
  BASE_LOCATION = 'npidata/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- FDA NDC Directory
CREATE ICEBERG TABLE ndc_directory
  BASE_LOCATION = 'ndc_directory/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- HCPCS Codes
CREATE ICEBERG TABLE hcpcs
  BASE_LOCATION = 'hcpcs/'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

Step 7: Query Your Data

-- Query NPPES provider directory
SELECT npi, provider_first_name, provider_last_name_legal_name,
       provider_business_practice_location_address_city_name,
       provider_business_practice_location_address_state_name
FROM mimi_ws_1.nppes.npidata
LIMIT 10;

Quick Reference: Complete Setup Script

Copy and customize this full script:

-- ============================================
-- MIMILABS SNOWFLAKE INTEGRATION SETUP
-- ============================================

-- 1. Create catalog integration (one-time)
CREATE OR REPLACE CATALOG INTEGRATION mimi_ws_1
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = DELTA
  ENABLED = TRUE;

-- 2. Create external volume
--    Replace placeholders with values from mimilabs
CREATE OR REPLACE EXTERNAL VOLUME mimilabs_volume
  STORAGE_LOCATIONS = (
    (
      NAME = 'mimilabs-data'
      STORAGE_PROVIDER = 'S3'
      STORAGE_BASE_URL = '{{S3_PATH}}'
      STORAGE_AWS_ROLE_ARN = '{{IAM_ROLE_ARN}}'
    )
  )
  ALLOW_WRITES = FALSE;

-- 3. Get credentials to send to mimilabs
DESC EXTERNAL VOLUME mimilabs_volume;
-- >>> Send STORAGE_AWS_IAM_USER_ARN and STORAGE_AWS_EXTERNAL_ID to mimilabs <<<

-- 4. Create database/schema (use mimi_ws_1 for mimibot query compatibility)
CREATE DATABASE IF NOT EXISTS mimi_ws_1;
CREATE SCHEMA IF NOT EXISTS mimi_ws_1.{{SCHEMA_NAME}};

-- 5. Create tables (after mimilabs confirms IAM setup)
CREATE ICEBERG TABLE mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}}
  BASE_LOCATION = '{{BASE_LOCATION}}'
  EXTERNAL_VOLUME = 'mimilabs_volume'
  CATALOG = 'mimi_ws_1'
  AUTO_REFRESH = TRUE;

-- 6. Query (same syntax as mimibot/Databricks)
SELECT * FROM mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}} LIMIT 10;

Refreshing Data

With AUTO_REFRESH = TRUE, your Iceberg tables automatically detect new data — no action required.

To manually trigger a refresh:

ALTER ICEBERG TABLE mimi_ws_1.{{SCHEMA_NAME}}.{{TABLE_NAME}} REFRESH;

To check the last refresh time:

SHOW ICEBERG TABLES LIKE '{{TABLE_NAME}}';

Troubleshooting

`not authorized to perform: sts:AssumeRole`

Cause: IAM trust policy not configured
Solution: Send us your STORAGE_AWS_IAM_USER_ARN and STORAGE_AWS_EXTERNAL_ID

`Failed to find the metadata of Delta table`

Cause: Wrong path or BASE_LOCATION
Solution: Verify BASE_LOCATION matches the table subdirectory

`Forbidden (Status Code: 403)`

Cause: IAM permissions issue
Solution: Contact mimilabs — we'll check the policy

`Invalid value for parameter 'CATALOG_SOURCE'`

Cause: Snowflake edition limitation
Solution: Delta Direct requires Snowflake Enterprise Edition or higher

Need Help?

Your Snowflake account identifier
Error messages (if any)
Which tables you're trying to access

❄️ Snowflake Integration Guide

Single Table Setup

Multiple Tables (Shared Parent Path)

not authorized to perform: sts:AssumeRole

Failed to find the metadata of Delta table

Forbidden (Status Code: 403)

Invalid value for parameter 'CATALOG_SOURCE'

Loading...

❄️ Snowflake Integration Guide

Single Table Setup

Multiple Tables (Shared Parent Path)

not authorized to perform: sts:AssumeRole

Failed to find the metadata of Delta table

Forbidden (Status Code: 403)

Invalid value for parameter 'CATALOG_SOURCE'

`not authorized to perform: sts:AssumeRole`

`Failed to find the metadata of Delta table`

`Forbidden (Status Code: 403)`

`Invalid value for parameter 'CATALOG_SOURCE'`

`not authorized to perform: sts:AssumeRole`

`Failed to find the metadata of Delta table`

`Forbidden (Status Code: 403)`

`Invalid value for parameter 'CATALOG_SOURCE'`