Greenplum or ClickHouse?

Question

SQL

Tell me, please, if we are talking about an analytical data bus

• dozens of columns, billions of rows
  
• data is not updated
  
• extremely rare, but voluminous requests (i.e. not from users, but from analysts)
  
• with good SQL support
  
• columnar
  
• more like a single server than a cluster
  
• other things being approximately equal, we choose the architecture easier
  

, then what is worth considering in 2020(1)?

Thank you!

PS In an article from 2016, it was stated that SQL support, for example, JOINs, is very limited in ClickHouse.
Is it relevant now?

PPS Greenplum or ClickHouse?

Answer 1

in terms of performance, ClickHouse is better than
JOIN, of course, since 2016 they have improved, but there are still a lot of limitations,
but for big data, JOIN of two large tables will still slow down
ClickHouse community is better IMHO
with GreenPlum, we don’t have many people with experience
if OLTP + OLAP workload is planned and cluster
, then you can generally look at TiDB and Citus