Documentation
¶
Overview ¶
Package mpp is what the analytical workloads share about Greenplum-family engines (Greenplum, Greengage, WarehousePG, Apache Cloudberry): the dialect name and the check that keeps a run on the dialect that matches the server it is connected to.
These engines speak the PostgreSQL protocol, so the pg driver runs them, and the pg schema loads and answers queries on them too. It is not the schema they are meant to run: without DISTRIBUTED BY every table is spread by its first column and every join moves data, so the numbers would be neither representative nor comparable. Neither a missing nor a wrong dialect is therefore ever accepted silently.
Index ¶
Constants ¶
const Dialect = "gp"
Dialect is the value of the sql-dialect parameter that selects the MPP schema.
Variables ¶
var ( ErrDriver = errors.New("sql-dialect needs the pg driver: Greenplum-family engines speak the PostgreSQL wire") ErrSchema = errors.New("sql-dialect gp brings its own schema; do not combine it with a schema override") ErrUnlog = errors.New("pg-unlogged does not apply to sql-dialect gp: Greenplum has no unlogged tables") ErrNeedGP = errors.New("the server is a Greenplum-family MPP engine: run with --sql-dialect gp " + "(the pg dialect declares no distribution or storage for its tables)") ErrNotGP = errors.New("--sql-dialect gp needs a Greenplum-family server: " + "this one has no pg_catalog.gp_segment_configuration") )
Functions ¶
This section is empty.
Types ¶
type Request ¶
type Request struct {
Dialect string
Driver bench.DriverTypeName
OwnSchema bool // the author supplies the schema SQL, so the built-in one is not used
Unlogged bool // pg-unlogged given
}
Request is what a workload knows when it picks its dialect.
type Selection ¶
type Selection struct {
// contains filtered or unexported fields
}
Selection is the dialect a run uses and, where it can be checked, the server check.