Open graft opened 6 years ago
@dabrau this is perhaps solveable with partitioning?
@graft partitioning will help if the query needs to access a subset of the child tables. So I think it will probably only make a difference in certain queries depending on how the table is partitioned.
The Magma /retrieve endpoint can stream data in TSV format; in order to do this Magma computes page bounds with one SQL query, then queries for records within those bounds with another SQL query. The latter query is always fast; the former, however, becomes extremely slow when retrieving records from a large table.