Data destinations

Data Exports supports a wide range of data destinations, allowing you to send your Analyze and Optimize data to your preferred data warehouse, database, or object storage solution.

Supported destinations

OLAP data warehouses

Cloud-based analytical data warehouses optimized for large-scale data analysis and reporting.

DestinationDescription
SnowflakeCloud data platform with near-unlimited scalability
Google BigQueryServerless, highly scalable data warehouse
Amazon RedshiftFully managed data warehouse service
Amazon Redshift ServerlessServerless option for Redshift
Amazon AthenaServerless query service for S3
DatabricksUnified analytics platform
ClickHouseColumn-oriented database for analytics
MotherDuckServerless analytics with DuckDB

Open table formats

Open-source table formats that provide ACID transactions on data lakes.

DestinationDescription
Delta LakeOpen-source storage layer with ACID transactions
Apache IcebergOpen table format for large analytic tables

OLTP databases

Transactional databases for operational workloads.

DestinationDescription
PostgreSQLPostgreSQL database
Aurora PostgreSQLAWS Aurora PostgreSQL
MySQLMySQL database
Aurora MySQLAWS Aurora MySQL
SQL ServerMicrosoft SQL Server
OracleOracle Database
MongoDBDocument database

Object storage

Cloud-based file storage for data lakes and archival.

DestinationDescription
Amazon S3AWS object storage
S3-compatible storageObject storage platforms that support S3 compatibility
Google Cloud StorageGCP object storage
Azure Blob StorageAzure object storage
SFTPSecure file transfer

Spreadsheets

Spreadsheet destinations for smaller datasets and reporting.

DestinationDescription
Google SheetsGoogle Sheets integration

Common configuration steps

Most destinations follow a similar setup process:

  1. Create a dedicated user - Create a special-purpose user in your destination to perform write operations
  2. Configure permissions - Grant the necessary permissions for the user to create schemas, tables, and write data
  3. Allowlist IP addresses - Add the required IP addresses to your destination’s firewall or security group
  4. Add credentials - Enter your destination’s connection details and credentials in the Data Exports settings

Format of landed data

Data warehouses and databases

Data transferred to data warehouses and relational databases is loaded as properly typed tables within a single schema. A _transfer_status table records transfer metadata, including a transfer_last_updated_at timestamp for each table.

Object storage

Data transferred to object storage destinations is loaded as Apache Parquet files (recommended) or CSV/JSON in Apache Hive-style partitions:

<bucket>/<folder>/<model>/dt=<date>/<part>_<timestamp>.parquet

Where:

  • <model> is the data model name (equivalent to a table name)
  • <date> is the transfer date (e.g., 2024-01-15)
  • <timestamp> is the transfer timestamp

Spreadsheets

Data transferred to spreadsheet destinations is loaded as a new tab per data model. Where possible, tabs are created as protected (read-only) to prevent accidental modification.

Next steps

Select your destination from the list above to view detailed configuration instructions.