gbase数据、南大通用产品文档：GBase8s

No current row
游标位置不正确。您必须首先使用 ResultSet.next() 、ResultSet.beforeFirst() 、
ResultSet.first() 或 ResultSet.absolute() 方法将游标放到结果集中。

High Performance
The single-node data computing engine automatically selects the optimal execution
plan and algorithm through rule-based and cost-based cost evaluation. Leveraging
columnar storage, efficient compression, and intelligent indexing technologies, it
saves three orders of magnitude in CPU and I/O resource consumption, and achieves
query performance 50 to 1000 times higher than traditional row-based databases.
It supports deploying multiple computing nodes on multi-NUMA node servers,
binding each computing node to a NUMA node to avoid performance degradation
caused by cross-NUMA node operations. Multi-instance deployment can improve
performance by more than 1x.
Massive Parallel Processing (MPP) Technology
GBase 8a MPP Cluster is an asymmetrically deployed federated architecture, where
the three core components can be deployed separately. The maximum number of
nodes supported for GCluster scheduling cluster and GCWare management is 64,
and the GNode computing cluster supports deployment of more than 1000 nodes,
capable of handling structured data of over 100 PB.

GBase 8a MPP Cluster Technical White Paper
- 14 -
General Data Technology Co., Ltd.

enable_hashagg
参数说明：控制优化器对Hash 聚集规划类型的使用。
该参数属于USERSET 类型参数，请参考表15-1 中对应设置方法进行设置。
取值范围：布尔型

on 表示使用。

off 表示不使用。
默认值：on