Table Columns

Since Amazon DynamoDB tables are schemaless, the adapter offers the following two mechanisms to uncover the schema.

Dynamic Schemas

The columns of a table are dynamically determined by scanning data in the first few rows. You can adjust the number of rows that are used by modifying the RowScanDepth property. In addition to the name of the column, the row scan also determines the data type. The following table shows how the different data types supported by Amazon DynamoDB are modeled in the adapter.

Amazon DynamoDB Type	Modeled Type	Encoding	Sample Value
Boolean	Boolean	Not Required	USA
String	String	Not Required	USA
Blob	String	Not Required
Number	Double	Not Required	24.0
String Array	String	JSON Array	["USA","Canada","UK"]
Number Array	String	JSON Array	[20,200.5,500]
Blob Array	JSON Array	JSON Array	["ABCD","EFGH"]
Document	JSON Object	JSON Object	{"Address":"123 Fake Street","City":"Chapel Hill","Zip":"27516"}
List	JSON Array	JSON Array	[{"S":"mystring"},{"NS":[1,2]},{"N":4}]

Static Schemas

Instead of using dynamically discovered schemas, you can define your own schemas. This will give you more control over the projected columns and also enable you to use other data types such as boolean, datetime, etc. Refer to the CreateSchema Stored Procedure in order to create your own schema. You can simply specify the FileName (fullpath) and TableName of the new schema file, which should match with the name of the Amazon DynamoDB table, and edit the column listing to use it for your own table.

Schemaless Operations

While the schema of the table is necessary to report metadata, data may be selected, inserted, updated, or deleted from columns that do not exist in the schema. Columns that do not already exist in the table schema will have their data types dynamically determined based on the data that is specified. See DynamoDB Queries for more information.