Let’s look at the definitions first. A data warehouse is said to be a combination of technologies and components that enables the storage and strategic use of data, which collects and manages from various sources to provide meaningful business information, based on numerous inquiries and various analyzes. In other words, it takes a large set of data and transforms it into valuable information for the organization.

Data lake instead, can store a large amount of structured, semi-structured and unstructured data, in its native format, without fixed limits in terms of size or type offering a large amount of data available for complex analysis. As a consequence, it accelerates business decision making, based on a clearer vision.

Initially, it could be said that both function as large storage repositories and that their objective is to allow the management of enormous volumes of data to obtain business responses. However, there are many differences that separate a data lake from data warehouses, such as:

– Flexibility. Data lake allows the storage of structured, unstructured and semi-structured data to be analyzed later together, regardless of the source. Data warehouse, by its nature, focuses on structured data.

– Architecture. While a data lake maintains data in its original format in a flat architecture, a data warehouse builds a hierarchical structure with folders and files. This makes the second one more agile to find and analyze the information.

– Adaptability. At a time of high dynamism, when the business needs to move at high speed, data lake offers such adaptability that it enables results faster than a data warehouse.

– Consumption. At the time it is going to be used, data lake transforms data into the model that is needed (in a format known as schema on read). Instead, data warehouse uses the schema model on write, outlining during writing, which promotes the fact that the data must fit into predefined schemas, which makes them less adaptable at the time of reading. This also enables the data from the data lake to be accessed in any situation, while data warehouses are restricted to providing information on predefined questions for predefined data types.

– Conservation. Data lake keeps all data, even data that does not have any specific use at an specifit time, but could have it in the future. Data warehouse, on the other hand, requires an effort of data profiling and decision-making of what is kept and what data should be excluded from the warehouse.

– Processing. While data warehouse uses the traditional ETL (Extract, Transform, Load), that is, it first extracts the data, then transforms it and finally makes it available to those who need it, data lake applies the ELT process (Extract, Load, Transform), speeding up the load to generate greater speed of response to queries.

– Typical user. Overall, data warehouse is the ideal format for operational users: the data is well structured and easy to use and understand. Meanwhile, data lake also accepts operational users, but also enables the possibility of more in-depth analysis, for which even data scientists applying advanced analytical tools may be necessary. However, for a less technical user data warehouse is simpler.

In short, data lake promotes greater agility, more flexibility, and greater depth than data warehouse in transforming data into value.