# Schema Creation

> Source: https://docs.synthesized.io/tdk/latest/user_guide/020_guides/managing_data_properties/schema_creation
> For the complete documentation index, see [llms.txt](https://docs.synthesized.io/llms.txt).

Synthesized can only insert data into the target database if the target schemas are compatible with the input schemas. Schema creation is the stage in the data processing pipeline responsible for modifying the target database schemas to match the source database.

While most database objects can be processed, some features must be manually configured for certain databases.

Synthesized offers various schema creation modes that allow you to handle your use cases.

## Prerequisites

Before running a Synthesized workflow for the first time on a new data source:

1. Check that the username Synthesized is using has the [appropriate access rights](https://docs.synthesized.io/tdk/latest/user_guide/010_get_started/database_permissions).
2. Check that the version of the two databases is the same.

If the Synthesized user does not have the appropriate permissions or there is a version mismatch, Synthesized may not be able to copy the schemas. In such cases, the database should be prepared manually and [`DO_NOT_CREATE`](#_drop_and_create) mode should be used.

## Best practices

Before running a Synthesized workflow for the first time, please check the following things:

1. If there is already data in the target database, create a backup.
2. If other workflows are generating data in the same target database, use either [`CREATE_IF_NOT_EXISTS`](#_create_if_not_exists) mode or [`DO_NOT_CREATE`](#_do_not_create) mode.

## Supported schema objects

Certain databases have functionality that is not able to be copied automatically. This table shows which features work with each database type.

      
| Feature | MSSQL | Postgres | Oracle | MySQL | SQLite | DB2 |
| --- | --- | --- | --- | --- | --- | --- |
| 
Schemas

 | 

 | 

 | 

N/A

 | 

 | 

N/A

 | 

 |
| 

Tables

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Indexes

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Constraints

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Sequences

 | 

 | 

 | 

 | 

 | 

N/A

 | 

N/A

 |
| 

Primary Keys

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Foreign Keys

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Triggers

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Domains

 | 

 | 

 | 

 | 

N/A

 | 

 | 

 |
| 

Views

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Custom Data Types

 | 

 | 

 | 

 | 

N/A

 | 

N/A

 | 

 |
| 

Computed Columns

 | 

 | 

 | 

 | 

 | 

 | 

 |
| 

Enums

 | 

N/A

 | 

 | 

N/A

 | 

N/A

 | 

N/A

 | 

N/A

 |

## Schema creation modes

To control how Synthesized prepares the target schema, use the `schema_creation_mode` flag. E.g.

By default, `DROP_AND_CREATE` is recommended, particularly when using Synthesized for `MASKING`. This ensures Synthesized is always operating on fresh schemas. However, this is only appropriate if all schema objects are supported.

In systems where you are generating partial data, `CREATE_IF_NOT_EXISTS` is recommended. This lets Synthesized create new schemas as they are required.

If you have manually prepared the database, `DO_NOT_CREATE` can be used to skip schema creation.

### DROP\_AND\_CREATE

1. Drops the existing schemas in the target database
2. Creates the new schemas
3. Fully recreates all objects

### CREATE\_IF\_NOT\_EXISTS

1. Calculates the difference between input and output schemas
2. Creates each schema if it doesn’t exist
3. Copies any missing schemas from input to output

### DO\_NOT\_CREATE

1. Does not create schemas in the target database
2. Verifies that required schemas already exist
3. Throws an exception if any required schema is missing

### CREATE

1. Verifies that the target database is empty
2. Creates a new schema in the target database
3. Throws an exception if objects already exist