Paul Dix
Paul Dix
Yeah, that's around what I was seeing, I think. I'm curious how it looks if you remove the segment info files for all but one segment and then restart the...
Going to close this now given all the changes and the profiling doc has been written up.
Yes, functionally they're very similar to tags. The critical differences are that ordering matters with series key columns, that series key columns can't be added later, and values can't be...
@jacobmarble what's the difference between structs and maps?
@jacobmarble on the timestamp question for how the user specifies that it's nanosecond when they write the data, that's why I mention that they can only do that with an...
@alamb I think maybe you're thinking that series key is one thing/value? That is not what I'm thinking. Series key is a logical concept, where it is an ordered set...
This isn't actionable at this point and we'll be revisiting data model later.
Partitions doesn't make sense to bring over as OSS doesn't use the same scheme. We'll want to have a system table for listing parquet files I think.
Yeah, the partition ID is just a stub to satisfy the trait. For the `system.parquet_files` table, I think it would be good to limit the scope of what the user...
@alamb I'm looking at the system tables as an API for users to find out information like schema, the parquet files that exist, and other information about the data itself....