Columnar databases do have their disadvantages. They are typically less efficient when it is necessary to update or delete data, for several reasons.
First and foremost, updating or deleting a single row of data requires finding several locations on disk where the individual columns are stored.
Even single row retrievals can be slower, resulting in a noticeable performance difference.
Columnar databases can also be implemented as MPP (massively parallel processing) systems, as hardware appliances or as in-memory systems.
First and foremost, updating or deleting a single row of data requires finding several locations on disk where the individual columns are stored.
Even single row retrievals can be slower, resulting in a noticeable performance difference.
Columnar databases can also be implemented as MPP (massively parallel processing) systems, as hardware appliances or as in-memory systems.
Comments