本文整理汇总了Golang中parser.QuerySpec.GetGroupByColumnCount方法的典型用法代码示例。如果您正苦于以下问题：Golang QuerySpec.GetGroupByColumnCount方法的具体用法？Golang QuerySpec.GetGroupByColumnCount怎么用？Golang QuerySpec.GetGroupByColumnCount使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类parser.QuerySpec的用法示例。
在下文中一共展示了QuerySpec.GetGroupByColumnCount方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Golang代码示例。
示例1: QueryResponseBufferSize
func (self *ShardData) QueryResponseBufferSize(querySpec *parser.QuerySpec, batchPointSize int) int {
groupByTime := querySpec.GetGroupByInterval()
if groupByTime == nil {
// If the group by time is nil, we shouldn't have to use a buffer since the shards should be queried sequentially.
// However, set this to something high just to be safe.
log.Debug("BUFFER SIZE: 1000")
return 1000
}
tickCount := int(self.shardNanoseconds / uint64(*groupByTime))
if tickCount < 10 {
tickCount = 100
} else if tickCount > 1000 {
// cap this because each response should have up to this number of points in it.
tickCount = tickCount / batchPointSize
// but make sure it's at least 1k
if tickCount < 1000 {
tickCount = 1000
}
}
columnCount := querySpec.GetGroupByColumnCount()
if columnCount > 1 {
// we don't really know the cardinality for any column up front. This is a just a multiplier so we'll see how this goes.
// each response can have many points, so having a buffer of the ticks * 100 should be safe, but we'll see.
tickCount = tickCount * 100
}
log.Debug("BUFFER SIZE: %d", tickCount)
return tickCount
}