Re-computation principle

Since raw data is always available in the lake, it's always possible to cater to new requirements by running or re-computing functions against the raw data. In addition to this, it's apt to store this data in a schema-less structure because tying data to a schema brings its own issue of re-computation. Tying data to schema also brings overhead to development and maintenance.
While implementing the Data Lake with this pattern as one of the main layers, we will see how these principles described before are realized.