Just finished watching a webinar of Snowflake’s Hybrid Tables. IMO, this is a HUUUUUUUGE deal. Companies can build applications, using row-store (good for lots of small transactions; think booking an airline or checking out at the grocery store). capabilities of it and it AUTOMATICALLY gets copied to a column-store (Good for analytical workloads & faster BIG queries). Snowflake under the hood automatically determines which version of the hybrid table to use to return the query.
The BIG BIG BIG benefit of using Snowflake is that it will be MUCH MUCH MUCH easier & cheaper to do analytics; no need to move data regularly and maintain these mega operations; especially when needing up-to-second accuracy on their reports. ETL processing is very time consuming and expensive. The most expensive and time consuming part of big reporting & analytics.
Use of Snowflake true cloud scale (NO other transactional database can scale CPU the way Snowflake can).
Can seamlessly join with traditional Snowflake tables & new Iceberg tables
Not ideal; won’t meet the needs for MEGA HIGH-SCALE transactional data loads (think IoT; monitoring flows etc…). Think 80/20 rule. 80% of workloads will run fine and not notice these types of issues. This is the subset of applications I think of as usually “developer preference”. They’ll run fine in SQL based (Oracle, SQL Server), NoSQL (MDB) etc…
Not yet fully featured (for example, triggers)