SAS data processing technology (1)

SAS数据处理技术

通常情况下,使用data stepDoing data processing is a good option. Data step for data processing,可以提供.

• Flexible programming capabilities
• Rich data processing functions

other tools and techniques

• SORT, SQL, 以及TRANSPOSEprocess step,It is very useful for data processing and transformation.

• SASThe macro function makes the code more flexible.

• 调试(Debugging)技术,Can be used to help identify logical errors.

SAS程序流程

Review some knowledge points

• SORTprocess step的OUT=Option to specify the creation of a new output dataset, rather than overwriting the original dataset.

• 用FORMATProcedures can create user-defined output formats and input formats.默认情况下,The created format will be stored in the directory

libname orion 's:\workshop';
data work.qtrlsalesrep;
proc sort data=work.qtrlsalesrep;

proc format;
value $ctryfmt 'AU'='Australia'
	       'US'='United tates';
run;

（选b; a是data那句,c是set那句）

• 由 INPUT语句和 赋值语句创建的变量会被重新初始化(Initialized to missing values).
• 由 SET语句Read variables are not reinitialized(will only be overwritten by subsequent data)

OUTPUTThe statement implements data transposition

Output multiple lines of observations

• 例中,Each row of records generates three observations.Each new observation will include a codenameIDand a test valueSCORE.

data A;
	input ID $ scorel-score3;
	drop scorel-score3; # Field names are not output(列名)
	score=scorel;output;
	score=score2;output;
	score=score3;output;
cards;
02126 99 96 94
02128 89 90 88
;
proc print;
run;

Use one data step to output create multipleSAS数据集

• in a data step,It can be done by converting the output dataset 名称列在DATAwithin a statement to create multiple datasets.
• 通过在OUTPUTSpecify the dataset name within the statement,Direct output to one or more specific datasets is possible

data usa australia other;
set orion.employee_addresses;
if Country='AU'then output australia;
else if Country='US'then output usa;
else output other;
run;

* 或者
data usa australia other;
	set orion.employee addresses;
	select (Country);
		when ('US')output usa;
		when ('AU')output australia;
		otherwise output other;
	end;
run;
*在SELECT语句中使用DO-END,当一个表达式为真时,使用DO-ENDStatements can execute multiple statements

Use conditional statements to control which observation is output to which dataset

• in the data step,用DROP语句和KEEP语句Controls which variables are output to the output dataset.

• 默认情况下,SASAll observations in the dataset are processed one by one.用FIRSTOBS=和OBS=Options control which observations are processed.

Create an accumulation variable

• The general form of a summation statement：

variable + expression;

• 求和语句

  • If the variable to the left of the plus sign did not exist before,create the variable
  • This variable is initialized to the value before the first loop of the data step0
  • This variable is automatically preserved
  • 执行时,Add the value of the expression to the variable
  • 忽略缺失值.

Cumulative summation of grouped data

• 定义First..和Last.过程
• Calculates the cumulative sum of grouped data
• Use subsetsIFstatement to output the specified observation

data deptsals(keep=Dept Deptsal);
	set SalSort;
	by Dept;
	if First.Dept then Deptsal=0;
	Deptsal+Salary;
	if Last.Dept;
run;

proc sort data=sashelp.cars
			out=cars;
	by Make;
run;

data total cars(keep=Make MSRP sum_price);
	set cars;
	by Make;
	if first.Make then sum_price=0;
	sum price+MSRP;
	if last.Make;
run;

proc print data=total_cars noobs;
	var Make sum_price;
	format sum price dollar10.2;
run;

Different types of data are read in different ways

列输入,Formatted input and list input are usedINPUTThere are three ways a statement reads data.

列输入(Column Input) Read in with column input,conditions that the data satisfies

• in fixed fields

• standard character or numeric values （如 58 -23 67.23 00.99 5.67E5 1.2E-2）

• INPUTThe general form of the statement input method ： INPUT variable <$> startcol-endcol...;

Read raw data files using formatted input method

• INPUTThe general form of the statement formatted input method： INPUT 指针控制 变量 输入格式...;
• Formatted input methods read data in the following ways：

  • Move the input pointer to the starting position
  • Name the variable
  • Specifies the input format
  • input @5 FirstName $10.;
  • Column control pointer：@n Move the pointer to n列 +n 将指针向后移动n列

Controls when records are loaded

• Read multiple records from the raw data file as one observation.

• DATA SAS-data-set;
  	INFILE 'raw-data-file-name';
  	INPUT specifications;
  	INPUT specifications;
  	<additional SAS statements>
  RUN;
  

• Read a raw data file of mixed-type records.

• The row pointer controller controls when new records are loaded

  • DATA SAS-data-set;
    	INFILE 'raw-data-file-name';
    	INPUT specifications/
    	specifications;
    	<additional SAS statements>
    RUN;
    

  • 当SAS遇到一个“/”时,The record for the next row is loaded. Line control pointer：# n 载入第n行 / Load the next line

• Read a subset of raw data files for mixed-type records.

Other tricks for the list input method

• Data at the end of the record are missing values.INFILE 'raw-data-file'MISSOVER;

• Missing data are represented by two consecutive delimitersINFILE 'file-name' DSD;.

• Each record contains multiple observations.INPUT var1 var2 var3 [email protected]@