SAS proc 制表，按特定变量排序
[英]SAS proc Tabulate, Order by specific variable
问题描述
我使用的是虚构数字，因为我无法发布原始数据集。 我有以下 SAS 代码：
PROC TABULATE DATA=table;
TITLE 'Offences per country and year;
CLASS year country offence_type;
TABLES country, year ALL, region /BOX="" MISSTEXT='0';
RUN;
这为每个国家/地区提供了一张这样的表格。
SWITZERLAND
year | theft | robbery | assault | …
------------------------------------
ALL | 1302 329 100 …
1980 | 321 100 24 …
1981 | 280 24 20 …
… | … … … …
现在，我想更改罪行类型/列的顺序。 最常见的犯罪类型（所有年份）应出现在左侧，不太常见的犯罪类型应出现在右侧（如上例所示）。
我已经在TABLES -statement 之后尝试了/ ORDER=FREQ ，但随后年份的顺序也发生了变化。
任何人都可以帮忙吗？ 谢谢！
1 个解决方案
解决方案1
1 2019-10-17 12:43:24
在使用ORDER=FREQ选项的单独CLASS语句中指定offence_type变量。
例子：
- 第一个表格显示 class 顺序是数据升序，
- 第二个表格显示了
CLASS <var>/ ORDER=FREQ的效果。 - 第三个表格显示了
BY <group-vars>的附加效果（以及一些避免标题高于每个BY组的技巧）。
ods listing;
options
pagesize=1000
nodate nonumber nocenter
formdlim=' '
formchar="|----|+|---+=|-/\<>*"
;
dm 'listing; clear';
data have(index=(country));
do country = 'Zeroland ', 'Switzerland', 'Crimeland';
do year = 1980 to year(today());
date = mdy(1,1,year);
do date = date to intnx('year',date,1);
crime = scan ('theft robbery assault parking', min(4,rand('geometric', 0.6)));
/***********************************************\
|* edit, make robbery predominant in Crimeland *|
\***********************************************/
if country = 'Crimeland' then do;
crime = scan ('robbery theft assault parking', min(4,rand('geometric', 0.6)));
end;
output;
end;
end;
end;
format date date9.;
run;
表格
proc tabulate data=have NOSEPS format=8.;
title1 "Tabulate: default class ordering Alphabetical";
*where country like 'S%' and year < 1984;
where year < 1982;
class country year ;
class crime;
tables country, ALL year, crime / condense;
run;
proc tabulate data=have NOSEPS format=8.;
title1 "Tabulate: CLASS / order=freq - crimes Frequency Descending";
title2 "freq ordering based on all data";
* where country like 'S%' and year < 1984;
where year < 1982;
class country year ;
class crime / order=freq;
tables ALL country, ALL year, crime / condense;
run;
data for_tabulate(index=(country));
set have;
title1 = 'NOBYLINE, TITLE in Data. / order=freq by group';
title2 = 'Country ' || country;
run;
options nobyline;
proc tabulate data=for_tabulate NOSEPS format=8.;
title;
* where country like 'S%' and year < 1984;
where year < 1982;
by country;
class title1 title2 year ;
class crime / order=freq;
tables
title1 * title2,
ALL year,
crime
/ condense
;
run;
清单 output
按字母顺序
Tabulate: default class ordering Alphabetical
country Crimeland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| |assault |parking |robbery | theft |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 69| 50| 438| 176|
|year | | | | |
|1980 | 39| 29| 214| 85|
|1981 | 30| 21| 224| 91|
------------------------------------------------------------
country Switzerland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| |assault |parking |robbery | theft |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 70| 44| 178| 441|
|year | | | | |
|1980 | 42| 27| 84| 214|
|1981 | 28| 17| 94| 227|
------------------------------------------------------------
country Zeroland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| |assault |parking |robbery | theft |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 74| 38| 189| 432|
|year | | | | |
|1980 | 36| 25| 89| 217|
|1981 | 38| 13| 100| 215|
------------------------------------------------------------
频率（总体）下降
Tabulate: CLASS / order=freq - crimes Frequency Descending
freq ordering based on all data
All
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 1049| 805| 213| 132|
|year | | | | |
|1980 | 516| 387| 117| 81|
|1981 | 533| 418| 96| 51|
------------------------------------------------------------
country Crimeland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 176| 438| 69| 50|
|year | | | | |
|1980 | 85| 214| 39| 29|
|1981 | 91| 224| 30| 21|
------------------------------------------------------------
country Switzerland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 441| 178| 70| 44|
|year | | | | |
|1980 | 214| 84| 42| 27|
|1981 | 227| 94| 28| 17|
------------------------------------------------------------
country Zeroland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 432| 189| 74| 38|
|year | | | | |
|1980 | 217| 89| 36| 25|
|1981 | 215| 100| 38| 13|
------------------------------------------------------------
频率（按组）下降
title1 NOBYLINE, TITLE in Data. / order=freq by group
and title2 Country Crimeland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| |robbery | theft |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 438| 176| 69| 50|
|year | | | | |
|1980 | 214| 85| 39| 29|
|1981 | 224| 91| 30| 21|
------------------------------------------------------------
title1 NOBYLINE, TITLE in Data. / order=freq by group
and title2 Country Switzerland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 441| 178| 70| 44|
|year | | | | |
|1980 | 214| 84| 42| 27|
|1981 | 227| 94| 28| 17|
------------------------------------------------------------
title1 NOBYLINE, TITLE in Data. / order=freq by group
and title2 Country Zeroland
------------------------------------------------------------
| | crime |
| |-----------------------------------|
| | theft |robbery |assault |parking |
| |--------+--------+--------+--------|
| | N | N | N | N |
|----------------------+--------+--------+--------+--------|
|All | 432| 189| 74| 38|
|year | | | | |
|1980 | 217| 89| 36| 25|
|1981 | 215| 100| 38| 13|
------------------------------------------------------------
如何使用SAS中的Proc制表符以百分比形式显示分析变量（0-1）？
[英]How to display analysis variable (0 - 1) as percentage using Proc Tabulate in SAS?
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.