Reading through the documentation and the nature of the data that I am dealing with, it looks like it's best to partition data by year-month for the last 12 months, and have a script that merges the ROS containers for the partitions that are before 12 months ago. (unless there's a better way to do it within Vertica itself)
Although everything seems to be easy to find my way in and out, I still have some questions about some statements in the online documentation:
1/ from the doc: https://my.vertica.com/docs/6.1.x/HTML/ ... #10444.htm
I hope that this meant to say "Vertica supports a max of 1024 ROS containers (not partitions)", because the way I am going to (dynamically) add partitions as data comes in, then merge at the tail would hit the 1024 partitions at some point, but should be within 13 ROS containers (12 for each prev. 12 months, and 1 for the rest)While HP Vertica supports a maximum of 1024 partitions, few, if any, organizations will need to approach that maximum.at maximum.
2/ from the doc: https://my.vertica.com/docs/6.1.x/HTML/ ... #14361.htm
Obviously, this is assuming that the TM.ActivePartitionCount is set to 1.By default, the Tuple Mover assumes that all loads and updates for partitioned tables are going to the same active partition. For example, if a table is partitioned by month, the Tuple Mover expects that after the start of a new month, no data is loaded into the partition for the prior month.
My question regarding this is; what happens if I have 1 active partition set, but I am inserting data for more than 1 partition? (I guess I am trying to understand what would happen if the above TM expectation isn't met)
Thank you for your help!