advantage:
-
Extremely high loading speed (the highest can be equal to the sum of all hard disk IO, which is basically the limit)
-
Suitable for large amounts of data rather than small data
-
Real-time loading of data is limited to adding (deleting and updating require decompressing Blocks, then computing and then recompressing and storing)
-
Efficient compression ratio saves not only storage space but also computing memory and CPU.
-
Great for aggregation operations.
shortcoming:
-
Not suitable for scanning small amounts of data
-
Not suitable for random updates
-
The batch update situation is different. Some well-optimized columnar databases (such as Vertica) perform better, and some databases that are not targeted for updates perform poorly.
-
Not suitable for real-time operations involving deletes and updates.