Soil stores its data on the local filesystem. The equivalent of the soil database instance in memory is a directory on a filesystem. This directory and everything beneath is controlled by soil.

So the basic association we need to do is to associate the in-memory represenation with the directory on the filesystem by doing

soil := Soil path: 'mydb'.

From here there are multiple options:

Creating a database

Before you can work with soil you need to create an instance.

soil := Soil path: 'mydb'.
soil initializeFilesystem 

The method #initializeFilesystem creates the directory mydb and creates some files/directories beneath it. This is the basic structure soil needs to operate. After calling #initializeFilesystem the variable soil holds a fully initialized and functional database instance.

The equivalent short form of the code above is

soil := Soil createOnPath: 'mydb'

Trying to create a database on a non-empty directory will throw an error. So soil will not overwrite a prior existing database instance

Opening a database

An existing soil instance can be opened with

soil := Soil path: 'mydb'.
soil open 

or with the short form

soil := Soil openOnPath: 'mydb'

Closing a database

When the soil instance is not used anymore it should be closed to free the open file streams and such. For this a simple

soil close

does that.

Deleting a database

You can delete soil in programmatical way using

soil destroy

this will delete the directory of the soil instance. For testing purposes it might be useful to start over with always a new instance. You can combine it this way

soil := (Soil path: 'mydb')
   destroy;
   initializeFilesystem.

Now every time this is executed it will delete the soil instance and recreate it from scratch