Data Warehouses and Data Lakes are storage repositories for data, used by enterprises to accumulate data from a wide range of sources. A data lake holds raw data in its native format, using a flat architecture. A data warehouse, in contrast, stores data hierarchically, in files or folders.
 

Fundamental Differences Between Data Warehouse and Data Lake

The fundamental difference between data warehouse and data lake is in the architecture. A data warehouse is highly-structured and integrated with the business process associated with the data. In contrast, data lakes are agile, with no rigid structure. The data lake holds large swathes of raw data in the native format of the data, including structured, unstructured, and semi-structured data. Data Lakes hold data under multiple nodes, and data which do not fit into a typical standardised data warehouse.

Data Warehouses adopt a schema-on-write approach. The data is modelled and structured at the time it is placed in the depository. Changing the structure is possible, but is a tedious and time-consuming process. Data Lakes, in contrast, adopts a schema-on-read approach. The user does not have to define the data structure or requirements until the data is actually accessed for any use, giving developers the freedom to configure the required query or data model as required. 
 

Application of Data Warehouse and Data Lakes

Data Lakes are useful in situations where access to raw data is required easily, such as when the business question is ambiguous, or the probable use case of the data cannot easily be predicted. However, when the business model or the use-case for the data is predictable, the structured approach of the data warehouse is more preferred. 

Of late, many businesses find Data Lakes an effective way to cope with uncertainties, and seize opportunities fast, in an extremely fluid business environment. The changing nature of data, with businesses being straddled with more complex, diversified, and varied data also make Data Lakes a preferred option to store and retrieve data. 

The fundamental differences notwithstanding, Data Warehouses and Data Lakes complement each other, and one is not necessarily better than the other. Enterprises may use the model which best suits their requirements.