The question of select a custom code (also known as custom code or manual code) compared to using a tool for ETL processes is the one you have to deal with every time you need to create a Data warehouse (DWH). Each has its advantages and disadvantages.
The manual code in the creation of a Data Warehouse
The alternative of custom code is a good solution when it comes to create a data warehouseBecause this option gives institutions the ability to code exactly what they want, expressed in how they would like their transformation programs to be structured.
Among the main advantages of using custom code, are the following:
- Low cost (they usually use the language of the house, as an example: C ++, PL / SQL, Java).
- The code is designed just for your needs.
- Program optimization is achieved.
- This option allows you to build what you want, when necessary.
Despite this, this system is not without its disadvantages. At the same time requiring a broad knowledge base for programmers, opting for manual encoding can have the following drawbacks:
- Efficiency in the administration and maintenance of programs is difficult to achieve.
- If there are any changes to the datastore, many programs could be affected.
- This option works without centralized code repository.
- Manual coding implies that metadata capabilities are limited.
- If the development cycle is longer.
- Debugging is more difficult.
- Audit capacity is limited.
Creating a data warehouse using an ETL tool
Wearing a ETL tool represents the other side of the charging equation. A ETL toolgenerally provides an easy-to-use interface for creating and managing transformations. In the same way as the previous answer, this method has its pros and cons, even when the advantages outweigh the disadvantages. The most important would be:
- Get a nice visual interface for creating and maintaining programs.
- Centralized program storage.
- The version control of the programs.
- The relative simplicity of transformation customization.
- The proper metadata support that you provide.
- Rapid implementation of transformations.
- The purification system integrated into most products.
- The opportunity to transform programming.
- The opportunity to transform auditing.
Faced with the multiple benefits of this option, there are some disadvantages like:
- The high initial cost involved.
- Limited user knowledge of most products.
- Optimization, which is sometimes limited due to generic programming methods.
Given the main characteristics of both options and the advantages they provide, and also taking into account its negative aspects, It seems obvious that the balance is tilting towards the ETL tools, that more than an alternative, are actually a necessity in any organization, yes not always, at least sometime in the life cycle of your data storage system.
