这个功能帮助您存储每个片段的数据分布。
This feature helps you to store data distributions per fragment.
在这种模式下，不会生成数据分布。
Slice内的数据分布策略。
统计信息显示出表中的数据分布状况。
不恰当的分区键可能导致数据分布不均匀。
An inappropriate partitioning key can cause uneven data distribution.
特定列中的数据分布可能随时间发生改变。
Data distribution in particular columns may change over time.
出现这类差异的原因包括统计相关和数据分布。
These differences can exist due to statistical correlation and data distribution, among other reasons.
优化器可以使用数据分布来计算查询中过滤器的选择率。
The optimizer can use data distributions to calculate selectivity for the filters in a query.
收集数据分布的统计信息，然后再次生成访问计划。
Collect distribution statistics and then generate the access plan again.
余下的参数与数据分布或复制策略中的参数具有同样的语义。
The rest of the parameters have the same semantics as in Data Distribution or Replication Policy.
持久化一个类型被复制过的实例会调用复制策略而非数据分布策略。
Persisting an instance whose type is replicated invokes a Replication Policy instead of a Data Distribution Policy.
但是，使用该方式计算的选择性不如使用数据分布计算的选择性准确。
But the selectivity calculated in this way is not as accurate as selectivity calculated using data distribution.
不过，报表应用程序应该不需要知道这种可通过联合方法提供的数据分布。
However, the reporting application should not need to be aware of this data distribution which can be provided through the federated approach.
添加新的容器时，将启动一个自动的重新均衡操作以便将数据分布到所有容器上。
When new containers are added, an automatic rebalancing starts to distribute the data across all containers.
传统的缓存解决方案为了规避这些问题，都将数据分布在缓存节点组成的集群上。
Traditionally, caching solutions have sought to avoid these issues by distributing the data over a cluster of caching nodes.
优化器使用表大小、索引可用性以及数据分布等信息来确定响应该查询的最佳路径。
The optimizer USES information such as table sizes, availability of indexes, and data distribution to figure out the optimal way to answer the query.
假定正常数据分布的开销评估可能并不能最好地反映查询使用不同参数时的实际开销。
A cost estimation that assumes normal data distribution may not best reflect the real cost of a query when different parameters are used.
用户可以通过使用数据库分区组和表空间，来确定将表数据分布到哪些数据库分区上。
Using database partition groups and tablespaces, a user can determine across which and how many database partitions their table data is to be spread.
然而在缺乏数据分布信息的情况下，优化器将基于表索引计算不同类型筛选器的选择性。
However, in the absence of data distribution information, the optimizer will calculate selectivity for filters of different types based on table indexes.
RUNSTATS收集两种类型的数据分布统计信息：频率统计信息和分位数统计信息。
RUNSTATS collects two types of data distribution statistics: frequency statistics and quantile statistics.
但是，如果没有数据分布，则数据库服务器根据表索引计算不同类型的过滤器的选择率。
However, in the absence of data distributions, the database server calculates selectivity for filters of different types based on table indexes.
并行计算要涉及到组合大量数据并将那些数据分布到每个计算结点，以进行更快的处理。
Parallel computing involves combining large amounts of data and spreading that data across each of the compute nodes for faster processing.
只要数据分布自上一次RUNSTATS以后没有改变，则优化器的估计可能非常接近实际值。
The optimizer's estimate could be very close to the actual value as long as the data distribution has not changed since the last RUNSTATS.
Jan提醒说这个技巧还可以用来实现其他的数据分布策略，例如分片(Sharding)。
Jan notes that the technique can also be used to implement other data distribution strategies, such as sharding.
ObjectGrid可以像使用内存内数据库一样使用，也可以使用其将数据分布在网络上。
ObjectGrid can be used simply as an in-memory database, or it can be used to distribute data across a network.
为了填充数据分布，以medium或high模式运行UPDATESTATISTICS。
To populate the distribution of data, run UPDATE STATISTICS in the medium or high mode.
优化器使用UPDATESTATISTICS所生成的数据分布信息来计算查询中筛选器的选择性。
The optimizer USES data distribution information generated by UPDATE STATISTICS to calculate selectivity for the filters in the query.
虽然输入参数的语义与数据分布策略内的相同，返回值现在包含一个片标识的数组，而非单一一个片标识。
While the semantics of the input arguments remain the same as in the Data Distribution Policy, the return value now contains an array of slice identifiers instead of a single one.
但当科学家们开始将这些数据分布在地图上观察时，他们发现了一些模糊的黑点，表示无线电频率的扰动。
But when scientists started looking at the data overlaid onto a map, they saw dark spots of fuzz, indicating radio frequency interference.
出于可伸缩性方面的原因(如允许非常大的数据集)，或仅仅因为弹性，您可能希望让数据分布在多个JVM上。
For scalability reasons, such as allowing for very large data sets, or simply for resilience, you might want to distribute the data across many JVMs.
应用推荐