Around 85% of all Data lake projects fail. These are horrible numbers and there are numerous reason why Data lake projects performs so bad.
One of the reason, in my opinion, is that organisation ignores the rule of importing data in its native format. But what is a native format and how would you define it? Exemple: Data that exists in a relational database of an operational system, if Data gets extracted to a flat file, is it still in its native format? What do you think?
I was under the impression a data lake is akin to a PSA (Persistent Staging Layer)… which is kind of hard to fail at since its just maintaining history of your source data. Our DV is built on top of our PSA.
Data lake is akin of psa but a psa is not a data lake. A data lake requires data to be mounted (schema on read) and is used to store big data.