SAS笔记

Reading data into SAS

1. Manual data entry with data step

data example1;
	input subject age sex $ totchol bmi cigpday cvd $;
	datalines;
1 39 M 195 26.97 0 Yes
2 46 F 250 28.73 0 No
3 48 M 245 25.34 20 .
4 61 F  225 28.58 30 Yes
;
run;

2. infile statement with data step (can read any delimited text file)

data subset1;
	infile "E:\SAS_video_1_2\framingham_subset.csv" delimiter="," firstobs=2;
	/* add a colon modifier before the data format. The colon prevents read errors and allows SAS to correctly handle missing leading zeros in one digit months and days */
	input RANDID INTAKE : MMDDYY10. TOTCHOL AGE SEX $ SYSBP;
	format INTAKE MMDDYY10.
run;

3. proc import statement

/* read Excel file */
proc import out=subset2
					datafile= "E:\SAS_video_1_2\framingham_subset.xlsx"
					dbms=XLSX REPLACE;
		/* grab data from the worksheet "subset" in the original Excel file*/
		sheet="subset"
		/* There's a header row in the Excel file and SAS should use it to assign the variable names */
		getnames=Yes;
run;

/* read csv file */
proc import out=subset3
					datafile= "E:\SAS_video_1_2\framingham_subset.csv"
					dbms=CSV REPLACE;
		getnames=Yes;
run;

4. import wizard

/* generate proc import code for reading*/

SAS procedures for viewing data

1. Summaries of properties of SAS data file proc contents

/* order=varnum: requests variable number ordering instead of alphabetical ordering in proc contents */
proc contents data=example1 order=varnum;
run;

2. View actual file proc print

/* print data */
proc print data=example1 order=varnum;
run;

/* print subset of observations and variables */
proc print data=example1 order=varnum (obs=5);
	/* don't need to add dollar sign here */
	var subject sex cvd;
run;

How to do statistical tests in sas:
Wilcoxon Rank sum: https://www.stat.purdue.edu/~tqin/system101/method/method_wilcoxon_rank_sum_sas.htm
Kruskal Wallis:
https://www.stat.purdue.edu/~tqin/system101/method/method_kruskal_wallis_sas.htm
One-way ANOVA:
https://www.stat.purdue.edu/~tqin/system101/method/method_one_way_ANOVA_sas.htm
t-test:
https://www.stat.purdue.edu/~tqin/system101/method/method_two_t_sas.htm