Hey folks! I want to know and understand while using photon acceleration, there is a feature called columnar acceleration which basically is a method of storing data in columns rather than rows, which is particularly advantageous for analytical databases and data warehouses. I want to know that how it actually works and what makes it efficient during the ETL process?

Where as Vectorized query processing operates on batches of rows (vectors) rather than processing each row individually but how it actually makes a difference, whether we read the data in batches of by rows individually, anyhow we are reading each row only, so what makes there two features efficient in the case of photon acceleration?