英语翻译The table below gives the value of these properties for

英语翻译
The table below gives the value of these properties for var-
ious encoding schemes.Note that there are many variations
of each scheme.For example,we experimented with three
versions of dictionary encoding before settling on the one
described in this paper; in one version there was a single
dictionary entry per row value – i.e.,a standard row-based
dictionary scheme; another version was a pure column-based
scheme that did not gracefully degenerate into single values
as in the current scheme.In most cases,each variation of
the same scheme will have the same block properties in the
table below.A no/yes entry in the table indicates that the
compression scheme is agnostic to the property and the value
is determined by the data.
When an operator cannot operate on compressed data (if,
for example,it cannot make any optimizations based on the
block properties),it repeatedly accesses the block through
an iterator,as described in Section 5.1.If,however,the
operator can operate on compressed data,it can use the
block information methods described in Section 5.1 to take
shortcuts in operation.For example,the pseudocode for a
Count aggregator is shown in Figure 2.Here,the passed
in column is used for grouping (e.g.,in a query of the form
SELECT c1,COUNT(*) FROM t GROUP BY c1).(Note:this
code is simplified from the actual aggregation code for ease
of exposition).

下表给出了这些特性的VaR值—各种编码方案.请注意,有许多变化各方案.例如,我们尝试用三字典编码解决之前的一个版本在本文中描述的；在一个版本有一个单一的字典条目每行值–即,一个标准的基于行的字典方案；另一个版本是一个纯粹的基于列方案不优雅地退化为单值
在当前的方案.在大多数情况下,每一个变化该方案将在同一块属性下表.没有/是的条目表中表明压缩方案的性能不可知论和价值通过测定的数据.当操作员不能使用压缩的数据（如果,例如,它不能根据任何优化块属性）,它通过多次访问块一个迭代器,如第5.1节所描述的.如果,然而,这运营商可以在压缩数据上操作,可以使用块的信息的方法在5.1节拿快捷键操作.例如,一个伪代码
计数器如图2所示.在这里,通过在柱用于分组（例如,在一个查询的形式选择C1,计数（*）从T组C1）.（注：这个代码是简化的实际代码易于聚集博览会）