129_Power PivotPower BI DAX不同维度动态展示动态坐标轴

博客：www.jiaopengzi.com

焦棚子的文章目录
请点击下载附件

一、背景

某天在和**那还是叫我大铁吧** 交流关于季度&月度同时展示的问题，感概中国式报表真的需求很微妙。

下面来看看到底是什么。

A、效果图

B、动态效果

[video width=“852” height=“608” mp4=“https://jiaopengzi.com/wp-content/uploads/2020/03/129_2.mp4”][/video]

接下来我们用三种度量的写法来说明。

方法3由**那还是叫我大铁吧** 提供，我们感谢他。

二、数据源

1、日期表：calendar

因为是日期时间维度的不同展现，都围绕着日期表做文章，所以建立一个合适的日期表很重要。

个人喜欢用pq生成日期表code如下：

letdate_start=#date(2018, 1, 1),//开始日期date_end=#date(2019, 12, 31),//结束日期count=Duration.Days(date_end-date_start),//间隔天数calendar = #table(type table[dates=date,year=Int64.Type,month=Int64.Type,day=Int64.Type,year_m=Int64.Type,quarter=Int64.Type,year_q=Int64.Type,weeknumber=Int64.Type,weekday=Int64.Type,weekus=text,weekcn=text,weekcna=text,M=text,YM=text,Q=text,YQ=text],List.Transform({0..count},(n)=> let d=Date.AddDays(date_start,n) in {d//日期,Date.Year(d)//年,Date.Month(d)//月,Date.Day(d)//天,Date.Year(d)*100+Date.Month(d)//年月组合,Date.QuarterOfYear(d)//季度,Date.Year(d)*100+Date.QuarterOfYear(d)//年季度组合,Date.WeekOfYear(d,1)//周一开始的一年第几周,Date.DayOfWeek(d,1)//星期几,Date.DayOfWeekName(d, "en-US")//星期几英文,Date.DayOfWeekName(d, "zh-CN")//星期几中文,Text.End(Date.DayOfWeekName(d, "zh-CN"),1)//星期几简写,"M"&Date.ToText(d,"MM"),Date.ToText(d,"Yyy")&"M"&Date.ToText(d,"MM"),"Q"&Number.ToText(Date.QuarterOfYear(d)),Date.ToText(d,"Yyy")&"Q"&Number.ToText(Date.QuarterOfYear(d))}))
incalendar

2、事实表：data

用pq生成的随机数据（拿到附件后，若数据和文章截图不一样是正常的。）

3、维度表

分别为事实表维度表，以及季度&月度辅助表。

4、关系视图

A、方法1&方法3的关系视图

B、方法2的关系视图

注意：方法2中两条虚线的使用。

三、上DAX

1、基础度量：total

total = SUM(data[value])

2、方法1：test0

test0 =
VAR N =MAX ( 'CALENDAR'[MONTH] )
VAR T1 =SUMMARIZE (FILTER ( ALL ( 'CALENDAR' ), 'CALENDAR'[MONTH] <= N ),'CALENDAR'[M],'CALENDAR'[Q],'CALENDAR'[MONTH],'CALENDAR'[QUARTER])
VAR T2 =ADDCOLUMNS ( T1, "@QM", IF( N / 3 <= [QUARTER], [M], [Q] ) )//是否显示整季度关键："<="
VAR T3 =SUMMARIZE ( T2, [@QM] )
VAR TQ =FILTER ( T3, LEN ( [@QM] ) <= 2 )
VAR TM =FILTER ( T3, LEN ( [@QM] ) > 2 )
VAR I =SUM ( 'aux0'[INDEX] )
RETURNSWITCH (TRUE (), I >= 1&& I <= 4, CALCULATE ('Measure'[total]//基础度量,TREATAS ( TQ, 'CALENDAR'[Q] ),TREATAS ( VALUES ( 'aux0'[QM] ), 'CALENDAR'[Q] ),ALL ( 'CALENDAR'[M] )),I >= 5&& I <= 16, CALCULATE ('Measure'[total]//基础度量,TREATAS ( TM, 'CALENDAR'[M] ),TREATAS ( VALUES ( 'aux0'[QM] ), 'CALENDAR'[M] ),ALL ( 'CALENDAR'[M] )),BLANK ())

3、方法1：test1

test1 =
VAR N =MAX ( 'CALENDAR'[MONTH] )
VAR T1 =SUMMARIZE (FILTER ( ALL ( 'CALENDAR' ), 'CALENDAR'[MONTH] <= N ),'CALENDAR'[M],'CALENDAR'[Q],'CALENDAR'[MONTH],'CALENDAR'[QUARTER])
VAR T2 =ADDCOLUMNS ( T1, "@QM", IF( N / 3 < [QUARTER], [M], [Q] ) )//是否显示整季度关键："<="
VAR T3 =SUMMARIZE ( T2, [@QM] )
VAR TQ =FILTER ( T3, LEN ( [@QM] ) <= 2 )
VAR TM =FILTER ( T3, LEN ( [@QM] ) > 2 )
VAR I =SUM ( 'aux0'[INDEX] )
RETURNSWITCH (TRUE (), I >= 1&& I <= 4, CALCULATE ('Measure'[total]//基础度量,TREATAS ( TQ, 'CALENDAR'[Q] ),TREATAS ( VALUES ( 'aux0'[QM] ), 'CALENDAR'[Q] ),ALL ( 'CALENDAR'[M] )),I >= 5&& I <= 16, CALCULATE ('Measure'[total]//基础度量,TREATAS ( TM, 'CALENDAR'[M] ),TREATAS ( VALUES ( 'aux0'[QM] ), 'CALENDAR'[M] ),ALL ( 'CALENDAR'[M] )),BLANK ())

4、方法2：test0

test0 =
VAR N =MAX ( 'CALENDAR'[MONTH] )
VAR IQ=MAX('calendar'[quarter])
VAR I =SUM ( 'aux0'[INDEX] )
RETURNSWITCH (TRUE (), I < IQ, CALCULATE ('Measure'[total]//基础度量,USERELATIONSHIP('calendar'[Q],'aux0'[QM]),ALL ( 'CALENDAR'[M] )),I<=IQ*3+4&& I >(IQ-1)*3+4&&I<=N+4, CALCULATE ('Measure'[total]//基础度量,USERELATIONSHIP('calendar'[M],'aux0'[QM]),ALL ( 'CALENDAR'[M] )),BLANK ())

5、方法2：test1

test1 =
VAR N =MAX ( 'CALENDAR'[MONTH] )
VAR IQ=MAX('calendar'[quarter])
VAR I =SUM ( 'aux0'[INDEX] )
VAR TF=MOD(N,3)
RETURNSWITCH (TRUE (), I <= IQ && NOT(TF), CALCULATE ('Measure'[total]//基础度量,USERELATIONSHIP('calendar'[Q],'aux0'[QM]),ALL ( 'CALENDAR'[M] )), I < IQ && TF, CALCULATE ('Measure'[total]//基础度量,USERELATIONSHIP('calendar'[Q],'aux0'[QM]),ALL ( 'CALENDAR'[M] )),I<IQ*3+4&& I >(IQ-1)*3+4 && TF, CALCULATE ('Measure'[total]//基础度量,USERELATIONSHIP('calendar'[M],'aux0'[QM]),ALL ( 'CALENDAR'[M] )),BLANK ())

6、方法3：test0

test0 =
VAR FindQ =FIND ( "Q", SELECTEDVALUE ( aux0[QM] ),, 0 ) //是否存在Q
VAR DQ =CALCULATE (MIN ( 'calendar'[Q] ),FILTER ( 'calendar', [Q] = SELECTEDVALUE ( 'calendar'[Q] ) ) ) //当前季度
VAR DM =CALCULATE (MIN ( 'calendar'[M] ),FILTER ( 'calendar', [M] = SELECTEDVALUE ( 'calendar'[M] ) )) //当前月份
VAR DQMinMonth =CALCULATE ( MIN ( 'calendar'[M] ), FILTER ( ALL ( 'calendar' ), [Q] = DQ ) ) //当前季度最小月份
VAR RQ =FILTER ( VALUES ( 'aux0'[QM] ), LEN ( [QM] ) = 2 )//季度
VAR RM =FILTER ( VALUES ( 'aux0'[QM] ), LEN ( [QM] ) > 2 )//月
RETURN
SWITCH (TRUE (),FINDQ > 0&& SELECTEDVALUE ( 'aux0'[QM] ) < DQ, CALCULATE (SUM ( DATA[VALUE] ),TREATAS ( RQ, 'calendar'[Q] ),TREATAS ( VALUES ( aux0[QM] ), 'calendar'[Q] ),ALL ( 'calendar'[M] )),FINDQ = 0&& SELECTEDVALUE ( aux0[QM] ) <= DM&& SELECTEDVALUE ( aux0[QM] ) >= DQMINMONTH, CALCULATE (SUM ( DATA[VALUE] ),TREATAS ( RM, 'calendar'[M] ),TREATAS ( VALUES ( aux0[QM] ), 'calendar'[M] ),ALL ( 'calendar'[M] )),BLANK ()
)</code></pre>

``7、方法3：test1

test1 =
VAR TF=MOD(AVERAGE('calendar'[month]),3)//季度月末显示季度or月度
VAR FindQ =FIND ( "Q", SELECTEDVALUE ( aux0[QM] ),, 0 ) //是否存在Q
VAR DQ =CALCULATE (MIN ( 'calendar'[Q] ),FILTER ( 'calendar', [Q] = SELECTEDVALUE ( 'calendar'[Q] ) ) ) //当前季度
VAR DM =CALCULATE (MIN ( 'calendar'[M] ),FILTER ( 'calendar', [M] = SELECTEDVALUE ( 'calendar'[M] ) )) //当前月份
VAR DQMinMonth =CALCULATE ( MIN ( 'calendar'[M] ), FILTER ( ALL ( 'calendar' ), [Q] = DQ ) ) //当前季度最小月份
VAR RQ =FILTER ( VALUES ( 'aux0'[QM] ), LEN ( [QM] ) = 2 )//季度
VAR RM =FILTER ( VALUES ( 'aux0'[QM] ), LEN ( [QM] ) > 2 )//月
RETURN
SWITCH (TRUE (),FINDQ > 0 && SELECTEDVALUE ( 'aux0'[QM] ) <= DQ&& NOT(TF), CALCULATE (SUM ( DATA[VALUE] ),TREATAS ( RQ, 'calendar'[Q] ),TREATAS ( VALUES ( aux0[QM] ), 'calendar'[Q] ),ALL ( 'calendar'[M] )),FINDQ > 0 && SELECTEDVALUE ( 'aux0'[QM] ) < DQ , CALCULATE (SUM ( DATA[VALUE] ),TREATAS ( RQ, 'calendar'[Q] ),TREATAS ( VALUES ( aux0[QM] ), 'calendar'[Q] ),ALL ( 'calendar'[M] )),FINDQ = 0&& SELECTEDVALUE ( aux0[QM] ) <= DM&& SELECTEDVALUE ( aux0[QM] ) >= DQMINMONTH && TF, CALCULATE (SUM ( DATA[VALUE] ),TREATAS ( RM, 'calendar'[M] ),TREATAS ( VALUES ( aux0[QM] ), 'calendar'[M] ),ALL ( 'calendar'[M] )),BLANK ()
)</code></pre>

四、总结

1、思维依然是业务转换是筛选上下文；

2、方法1&方法3都是利用TREATAS创建虚拟关系来实现目的；

3、方法2中利用的虚线关系USERELATIONSHIP实现目的，同时方式2兼容PP，目前TREATAS不兼容PP，（更新：office365支持TREATAS）

4、本案例可以衍生如：动态坐标轴、动态维度展示等，可以举一反三的。

彩蛋：季度&月度动态坐标轴

相对比较简单就不单独赘述；见文章素材附件。

[video width=“1064” height=“924” mp4=“https://jiaopengzi.com/wp-content/uploads/2020/03/129_9.mp4”][/video]

by 焦棚子

焦棚子的文章目录

` ``