As shown in the IXSCAN descriptor in Figure 8, the leading column of the index exploited by optimizer is PRODUCT_NUMBER, which does not provide the order required for merge scan join.
如图8的IXSCAN描述符所示,通过优化器利用的索引的领先列是PRODUCT _ NUMBER,其不为合并扫描联接提供所需的顺序。
For the merge scan join to perform properly, both inputs are expected to be in the order of join columns.
为了合并扫描联接正确地执行,按照联接列的顺序预计有两个输入。
If your timestamps are to the nanosecond, you can re-merge these files together and usually not lose the order of events ("sort -m").
如果你的时间戳精确到纳秒,你可以重新合并这些文件并且通常不会打乱事项的顺序(“sort -m”)。
We said each of the merge operations O was of order n. But n is different. Right?
注意这里发生了什么,我们说过每一次合并操作的复杂度都是?
应用推荐