On-stack developer extensibility is a type of extensibility that allows you to create development projects directly on the SAP S/4HANA Cloud technology stack. It gives you the opportunity to develop cloud-ready and upgrade-stable custom ABAP applications and services inside the SAP S/4HANA Cloud, public edition system. You can use the ABAP Development Tools in Eclipse to create and deploy your on-stack extensions. On-stack developer extensibility is suitable for the following kinds of applications:
Applications that provide APIs for side by side SAP BTP apps. On-stack developer extensibility allows you to create OData services or RESTful APIs based on CDS view entities or projection views. These services or APIs can expose SAP S/4HANA data and logic to other applications that run on the SAP Business Technology Platform (SAP BTP) or other platforms. This way, you can create a loosely coupled integration between your SAP S/4HANA system and your side by side SAP BTP apps.
Applications that access SAP S/4HANA data using complex SQL. On-stack developer extensibility allows you to use ABAP SQL to access SAP S/4HANA data using complex queries, such as joins, aggregations, filters, parameters, and code pushdown techniques. You can also use ABAP SQL to perform data manipulation operations, such as insert, update, delete, and upsert. This way, you can create applications that require advanced data processing and analysis on SAP S/4HANA data.
The other kinds of applications are not suitable for on-stack developer extensibility, as they have different requirements and challenges. These kinds of applications are:
Applications that integrate data from several different systems. On-stack developer extensibility is not meant for creating applications that integrate data from multiple sources, such as other SAP systems, third-party systems, or cloud services. This is because on-stack developer extensibility does not support remote access or data replication, and it may cause performance or security issues. For this kind of applications, you should use side by side extensibility, which allows you to create applications that run on the SAP BTP and communicate with the SAP S/4HANA system via public APIs or events.
Applications that run separate from SAP S/4HANA. On-stack developer extensibility is not meant for creating applications that run independently from the SAP S/4HANA system, such as standalone apps, microservices, or web apps. This is because on-stack developer extensibility requires a tight coupling with the SAP S/4HANA system, and it may limit the scalability, flexibility, and portability of the applications. For this kind of applications, you should use side by side extensibility, which allows you to create applications that run on the SAP BTP and leverage the cloud-native features and services of the platform.

The access to internal tables can be optimized by using the appropriate table type and specifying the table key. The table key is a set of fields that uniquely identifies a row in the table and determines the sorting order of the table. The table key can be either the primary key or a secondary key. The primary key is defined by the table type and the table definition, while the secondary key is defined by the user using the KEY statement1.
The following results in faster access to internal tables:
B . In a sorted internal table, specifying the primary key completely. A sorted internal table is a table type that maintains a predefined sorting order, which is defined by the primary key in the table definition. The primary key can be either unique or non-unique. A sorted internal table can be accessed using the primary key or the table index. The access using the primary key is faster than the access using the table index, because the system can use a binary search algorithm to find the row. However, the primary key must be specified completely, meaning that all the fields of the primary key must be given in the correct order and without gaps2.
D . In a hashed internal table, specifying the primary key partially from the left without gaps. A hashed internal table is a table type that does not have a predefined sorting order, but uses a hash algorithm to store and access the rows. The primary key of a hashed internal table must be unique and cannot be changed. A hashed internal table can only be accessed using the primary key, not the table index. The access using the primary key is very fast, because the system can directly calculate the position of the row using the hash algorithm. The primary key can be specified partially from the left without gaps, meaning that some of the fields of the primary key can be omitted, as long as they are the rightmost fields and there are no gaps between the specified fields.
E . In a hashed internal table, specifying the primary key completely. A hashed internal table is a table type that does not have a predefined sorting order, but uses a hash algorithm to store and access the rows. The primary key of a hashed internal table must be unique and cannot be changed. A hashed internal table can only be accessed using the primary key, not the table index. The access using the primary key is very fast, because the system can directly calculate the position of the row using the hash algorithm. The primary key can be specified completely, meaning that all the fields of the primary key must be given in the correct order.
The following do not result in faster access to internal tables, because:
A . In a sorted internal table, specifying the primary key partially from the left without gaps. A sorted internal table is a table type that maintains a predefined sorting order, which is defined by the primary key in the table definition. The primary key can be either unique or non-unique. A sorted internal table can be accessed using the primary key or the table index. The access using the primary key is faster than the access using the table index, because the system can use a binary search algorithm to find the row. However, the primary key must be specified completely, meaning that all the fields of the primary key must be given in the correct order and without gaps. If the primary key is specified partially from the left without gaps, the system cannot use the binary search algorithm and has to perform a linear search, which is slower2.
C . In a standard internal table, specifying the primary key partially from the left without gaps. A standard internal table is a table type that does not have a predefined sorting order, but uses a sequential storage and access of the rows. The primary key of a standard internal table is the standard key, which consists of all the fields of the table row in the order in which they are defined. A standard internal table can be accessed using the primary key or the table index. The access using the primary key is slower than the access using the table index, because the system has to perform a linear search to find the row. The primary key can be specified partially from the left without gaps, but this does not improve the access speed, because the system still has to perform a linear search.