Data connector summary

Tonic Structural currently provides the following data connectors:

Connector	Type	Supported versions	Structural limitations and differences
Amazon Redshift	Data warehouse		Not available on Structural Cloud. No Incremental table mode. Some generators not supported. No upsert. No output to a container repository. Supports filters on de-identified tables.
Apache Iceberg	Data warehouse		No Incremental table mode. No subsetting No output to a container repository. No upsert. Post-job scripts require DuckDB. Supports filters on de-identified tables.
Databricks	Spark	9.1 - 14.3 LTS	No workspace inheritance. Limited to De-Identify and Truncate table modes. Some generators not supported. No subsetting. No output to a container repository. Supports table filters and partition settings on de-identified tables. No post-job scripts.
Db2 for LUW	Application database	V11.5	Enterprise licenses only. Not available on Structural Cloud. Limited to De-Identify and Truncate table modes. No upsert. No output to a container repository.
File connector	File connector	Can select .csv, .tsv, .json, .xml, and Parquet files from either Amazon S3, Google Cloud Storage, a local file system, or a file mount. Can also select .txt or .gzip files that contain CSV, XML, or JSON content.	Limited to De-Identify and Truncate table modes. No workspace inheritance. No subsetting. No upsert. No output to a container repository. No post-job scripts.
Google BigQuery	Data warehouse However, within Tonic it functions more like an application database. Tonic uses the REST API or Storage API to interact with the data.		No hidden datasets as source or destination. No Incremental table mode. No Cross-Table Sum generator. No subsetting. No upsert. No output to a container repository. Supports filters on de-identified tables.
MongoDB	Application database	7 and above Also supports Amazon DocumentDB	On the Structural application, the term "table" is generally replaced with "collection". Database View and Table View are replaced with Collection View. No Privacy Report download Limited to De-Identify, Truncate, and Preserve Destination modes for collections. Some generators not supported. Self-consistency only. No upsert. No output to a container repository. No post-job scripts. No workspace inheritance.
MySQL	Application database	8.0, 8.4, and 9.5 Also supports MariaDB 10.0 and above Amazon Aurora
Oracle	Application database	19c and above	Enterprise licenses only. No Incremental table mode. No output to a container repository.
PostgreSQL	Application database	14 through 18 Also supports Amazon Aurora
Salesforce	SaaS (Software as a Service)	Minimum Professional edition	Limited to De-identify and Truncate table modes. Some generators not supported. No upsert. No output to a container repository. No post-job scripts.
Snowflake	Data warehouse		Some generators not supported. No Incremental table mode. No upsert. No output to a container repository. Supports filters on de-identified tables.
Spark SDK (self-managed cluster)	Spark		Not available on Structural Cloud. No workspace inheritance. Limited to De-Identify and Truncate table modes. Some generators not supported. No subsetting. No upsert. No output to a container repository.
SQL Server	Application database	2016 and above Also supports Azure SQL