Is there an efficient way to get the latest rows (based on timestamp and we want all the columns in the row) grouped by a column? I tried using LASTWITHTIME but it only takes in 1 column name at a time. Specifying LASTWITHTIME for each of the columns in the query increases latency significantly, which seems to imply that it does a linear scan for each column?

can you please file a github issue with the details - sample queries, schema and latency numbers. There is a possibility to optimize this heavily based on the metadata