Go is the de facto standard in terms of programming in Microservice environments. Besides many other languages like Java, C#, Perl, Python, Ruby, Go is gaining popularity (see https://www.tiobe.com/tiobe-index/).
To be part of the Go community, all data stores must provide a Go-based interface to access data. The Adabas API for Go provides the possibility to maintain and handle Adabas metadata. Besides Java, Natural and 3GL languages it is compulsory and a logical way to have a state-of-the-art Go based interface for Adabas to be considered as a data store that can be easily used in projects.
The new interface addresses developers and software designers who work primarily with Go and related concepts. However, developers who are used to working with other programming languages also can benefit from this approach. The usage is similar to the Adabas Client for Java product that Software AG provides to work in Java environments.
The Adabas API for Go
- makes it easy for everyone to write Go-based applications and interact with Adabas databases.
- uses Go methods for data access.
Before diving deeper into technology let's explore the key points and the benefits. The intention is to offer an easy-to-use Adabas interface covering all functions of the native Adabas call interface, which is very powerful but hard to code. Of course, it is possible to use the Adabas call interface based on different raw buffer types in a Go program. A much more sophisticated way is to take the higher-level Adabas API using metadata to access the Adabas database.
Another important point is usability in general, but especially for developers who are not familiar with Adabas. Using the interface makes it easy to write applications based on data stored in Adabas. This is not only a topic for Linux, Unix, Windows (LUW) based organizations but also for companies running on mainframe e.g. z/OS.
Performance is always a topic. The interface is designed to provide optimal throughput and performance of the resources while processing Adabas data and communicating with Adabas. A positive side-effect is that the interface allows an easy integration of Adabas into other new technology like Microservices.
Actually, although it's notcompulsory for an interface, but worth mentioning, the interface fully supports Adabas transactions and the entire set of access methods and currently almost all of the special descriptor types of Adabas.
Software AG does not guarantee support if there are issues with the Adabas API for Go. All current Adabas database versions are accessible. On Adabas Mainframe there might be some issues. Current features include:
- Read, insert, update and delete records. Reference data in Adabas fields with Adabas short name or logical field names defined in Adabas maps
- Logical read using Adabas search
- Read in Adabas file physical order
- Read in Adabas file using the ISN order
- Read and Update "periodic groups" and "multiple fields". You can read and update the complete record not one specific index entry
- Read Adabas large objects but without partial reads
- Support for accessing databases:
- Using the native local Adabas client connection
- Implements the new Adabas TCP/IP direct link to the database
- Provision to set various Go log frameworks to trace operations
The graphic below shows how the Adabas API for Go works in general.
The client interface provides a set of functions which can be used in Go programs. A detailed description is available here. There are components to parse the queries and build internal calls based on the Adabas native interface. The transformed requests are sent to Adabas using the normal low level interface, the Adabas link module. This in turn allows including Entire Net-Work as middleware to reach remote databases. Ideally, the software runs local to the database. However, to reach a mainframe database Entire Net-Work is required.
Included in the sources is a test package that provides a selection of example Go programs that show the usage. All tests are based on the Adabas demo file Employees.
Most customers using Adabas have worked with Natural in the past. An important matter when migrating an application is the possibility to use Natural or Predict logical views. Adabas only allows using Adabas short name field definitions and the database ID. To provide a logical name of a database file and a logical name reference to the Adabas short name, a new Adabas Map is introduced. This Adabas Map will be stored inside the Adabas database in an extra Adabas repository file. Tools managing the Adabas Maps are part of the Adabas Client for Java product Software AG provides.
The new Adabas Map technology is needed because the Natural and Predict logical views are stored separate from the Adabas database. In order to use the logical view independent of Natural storage in Adabas is elementary. The benefit is that the logical view and the data can now be part of a backup strategy and multiple logical views, dependent on the work case, can be provided.
The following import file formats are supported:
- Natural SYSTRANS files including DDM definitions exported by Natural
- Remarks in the file definition files (FDT) can be used to define long names
- XML format defining the logical view