【学习】七天搞定SAS（三）：基本模块调用（格式、计数、概要统计、排序等）（上）

搞定基本的函数之后，开始鼓捣SAS里面的模型。也就是说，要开始写PROC了。说实话，越学SAS，越觉得SAS像Stata…无论是从输出 的样式，还是语法。好不习惯没有()的模型调用呀。若是说SAS和Stata的区别，怕只是Stata更侧重于计量模型而SAS则是服务于大多数统计模型 吧。

PROC的基本内容：CONTENT

先是一个最基本的PROC：content，可以显示数据集的主要特性。比如：LIBNAME tropical 'c:MySASLib'; PROC CONTENTS DATA = tropical.banana;

这里主要是两个声明:TITLE和FOOTNOTE。前者输出时候会产生一个标题，后者会产生尾注。用法也是比较直接的：TITLE ”Here’s another title”; TITLE ’Here’’s another title’; FOOTNOTE3 ’This is the third footnote’;

最后还有一个很像Stata的LABEL声明：LABEL ReceiveDate = ’Date order was received’ ShipDate = ’Date merchandise was shipped’;

可以变量加注释。其实R里面给变量加注释是一件非常麻烦的事情，只有少数几个包可以搞定，还非常不值的。一般说来，我尽量在变量命名的时候长一点，这样直接可以读懂；再就是重建一个新的表，存储变量名和label。

SAS PROC求子集:WHERE

如果要在PROC里面先求子集的话，可以直接调用WHERE。感觉这里和SQL的思路比较像。用法也算是比较简单（SAS里面的用法都不是很麻烦，除了某些模型）：PROC PRINT DATA = 'c:MySASLibstyle'; WHERE Genre = 'Impressionism'; TITLE 'Major Impressionist Painters'; FOOTNOTE 'F = France N = Netherlands U = US'; RUN;

这样最终得到的结果就是：Major Impressionist Painters 1 Obs Name Genre Origin 1 Mary Cassatt Impressionism U 3 Edgar Degas Impressionism F 5 Claude Monet Impressionism F 6 Pierre Auguste Renoir Impressionism F F = France N = Netherlands U = US

SAS PROC 数据进行排序:SORT

排序就更简单了，直接PROC SORT就可以了。DATA marine; INFILE 'c:MyRawDataLengths.dat'; INPUT Name $ Family $ Length @@; RUN; * Sort the data; PROC SORT DATA = marine OUT = seasort NODUPKEY; BY Family DESCENDING Length; PROC PRINT DATA = seasort; TITLE 'Whales and Sharks'; RUN;

这样数据就按照Family、Length（递减）排序了。Whales and Sharks 1 Obs Name Family Length 1 humpback 50.0 2 whale shark 40.0 3 basking shark 30.0 4 mako shark 12.0 5 dwarf shark 0.5 6 blue whale 100.0 7 sperm whale 60.0 8 gray whale 50.0 9 killer whale 30.0 10 beluga whale 15.0

SAS PROC 输出数据:PRINT

最简单的数据输出怕就是PRINT了，顾名思义，直接打印数据出来。这里可以进行便啦的选择，还就可以选择统计量：DATA sales; INFILE 'c:MyRawDataCandy.dat'; INPUT Name $ 1-11 Class @15 DateReturned MMDDYY10. CandyType $ Quantity; Profit = Quantity * 1.25; PROC SORT DATA = sales; BY Class; PROC PRINT DATA = sales; BY Class; SUM Profit; VAR Name DateReturned CandyType Profit; TITLE 'Candy Sales for Field Trip by Class'; RUN;

得到的结果为：Candy Sales for Field Trip by Class 1 ——————————– Class=14 ——————————— Date Candy Obs Name Returned Type Profit 1 Nathan 17612 CD 23.75 2 Matthew 17612 CD 17.50 3 Claire 17613 CD 13.75 4 Chris 17616 CD 7.50 5 Stephen 17616 CD 12.50 —– —— Class 75.00 ——————————– Class=21 ——————————— Date Candy Obs Name Returned Type Profit 6 Adriana 17612 MP 8.75 7 Caitlin 17615 CD 11.25 8 Ian 17615 MP 22.50 9 Anthony 17616 MP 16.25 10 Erika 17616 MP 21.25 —– —— Class 80.00 ====== 155.00

SAS PROC里面改变输出格式：FORMAT

基本就是FORMAT一下就可以了，再就是PUT的时候也可以调整。DATA sales; INFILE 'c:MyRawDataCandy.dat'; INPUT Name $ 1-11 Class @15 DateReturned MMDDYY10. CandyType $ Quantity; Profit = Quantity * 1.25; PROC PRINT DATA = sales; VAR Name DateReturned CandyType Profit; FORMAT DateReturned DATE9. Profit DOLLAR6.2; TITLE 'Candy Sale Data Using Formats'; RUN;

输出结果为：Candy Sale Data Using Formats 1 Date Candy Obs Name Returned Type Profit 1 Adriana 21MAR2008 MP $8.75 2 Nathan 21MAR2008 CD $23.75 3 Matthew 21MAR2008 CD $17.50 4 Claire 22MAR2008 CD $13.75 5 Caitlin 24MAR2008 CD $11.25 6 Ian 24MAR2008 MP $22.50 7 Chris 25MAR2008 CD $7.50 8 Anthony 25MAR2008 MP $16.25 9 Stephen 25MAR2008 CD $12.50 10 Erika 25MAR2008 MP $21.25

常用的格式有：

• 文本型：$HEXw.和$w.

• 日期型：DATEw.（输出为ddmmyy或者ddmmyyyy）、DATETIMEw.d（输出为ddmmyy:hh:mm:ss）、 DAYw.（输出为dd）、EURDFDDw. 、JULIANw.、MMDDYYw.（输出为mmddyy或mmddyyyy）、TIMEw.d（输出为hh:mm:ss）、WEEKDATEw.（输 出为工作日）、WORDDATEw.(输出为单词)。

• 数字型：BESTw.（自动选择）、COMMAw.d（逗号分隔）、DOLLARw.d（货币）、Ew.（科学计数法）、PDw.d、w.d（标准小数）。

输出的样本见下。

当然FORMAT还可以自定义factor型变量的输出格式，比如：DATA carsurvey; INFILE 'c:MyRawDataCars.dat'; INPUT Age Sex Income Color $; PROC FORMAT; VALUE gender 1 = 'Male' 2 = 'Female'; VALUE agegroup 13 -< 20 = 'Teen' 20 -< 65 = 'Adult' 65 – HIGH = 'Senior'; VALUE $col 'W' = 'Moon White' 'B' = 'Sky Blue' 'Y' = 'Sunburst Yellow' 'G' = 'Rain Cloud Gray'; * Print data using user-defined and standard (DOLLAR8.) formats; PROC PRINT DATA = carsurvey; FORMAT Sex gender. Age agegroup. Color $col. Income DOLLAR8.; TITLE 'Survey Results Printed with User-Defined Formats'; RUN;

就可以把数字型的1，2转换为对应的文本male和female等，还可以把变量离散化，得到的输出为：Survey Results Printed with User-Defined Formats 1 Obs Age Sex Income Color 1 Teen Male $14,000 Sunburst Yellow 2 Adult Male $65,000 Rain Cloud Gray 3 Senior Female $35,000 Sky Blue 4 Adult Male $44,000 Sunburst Yellow 5 Adult Female $83,000 Moon White

最终可以实现的自定义输出还包括简单的文本连接，比如：* Write a report with FILE and PUT statements; DATA _NULL_; INFILE 'c:MyRawDataCandy.dat'; INPUT Name $ 1-11 Class @15 DateReturned MMDDYY10. CandyType $ Quantity; Profit = Quantity * 1.25; FILE 'c:MyRawDataStudent.txt' PRINT; TITLE; PUT @5 'Candy sales report for ' Name 'from classroom ' Class // @5 'Congratulations! You sold ' Quantity 'boxes of candy' / @5 'and earned ' Profit DOLLAR6.2 ' for our field trip.'; PUT _PAGE_; RUN;

可以给出若干连续的输出（注意DATA _NULL_;将不生成任何SAS的数据表）：Candy sales report for Adriana from classroom 21 Congratulations! You sold 7 boxes of candy and earned $8.75 for our field trip. ———— Candy sales report for Nathan from classroom 14 Congratulations! You sold 19 boxes of candy and earned $23.75 for our field trip. ———— Candy sales report for Matthew from classroom 14 Congratulations! You sold 14 boxes of candy and earned $17.50 for our field trip. ————

原文始发于微信公众号（PPV课数据科学社区）：【学习】七天搞定SAS（三）：基本模块调用（格式、计数、概要统计、排序等）（上）