/* If-then statements and data merging.  Anne Fairlie.  October 2, 2008.  
For Graduate Lab Course*/

/*website to retrieve data file
'http://www.uri.edu/faculty/walls/solomon1.csv'   */


/*second try--reading in as comma delimited with input of real col names*/
filename sol url 'http://www.uri.edu/faculty/walls/solomon1.csv' 
/*user='Ted' /*prompt proxy='http://inetgw.unx.sas.com:80'*/;
data temp1;
infile sol dlm=',';
input ID warmcold convers generous happy goodnat humor sociable popular;
   firstobs = 2;
run; 

proc contents data = temp1; run;
data temp2; set temp1;
if id = . then delete;
proc print data = temp2; run;
proc univariate data=temp2;
var convers; run;


/*reading data in directly*/

data made_from_log_file;
   infile datalines delimiter=',';
   input ID warmcold convers generous happy goodnat humor sociable popular;
   datalines;
8,0,3,5,4,5,6,5,5
9,0,2,4,4,4,5,5,5
10,0,2,5,6,5,5,3,3
11,0,3,5,3,2,2,4,3
12,0,3,3,3,2,5,4,4
13,0,3,4,4,5,5,5,5
14,0,2,6,5,4,6,6,6
15,0,2,4,5,5,5,5,5
17,0,3,3,3,1,3,5,5
18,0,4,3,3,3,2,3,4
21,0,3,4,4,3,5,6,4
23,0,4,4,3,4,4,3,4
24,0,2,3,6,5,5,5,4
27,0,4,3,3,5,4,4,4
28,0,2,1,2,2,5,4,3
31,0,3,4,4,4,4,3,3
34,0,4,3,5,5,3,6,6
35,0,4,3,5,5,3,6,6
36,0,4,3,5,2,2,2,4
38,0,4,5,3,3,3,4,4
1,1,4,2,3,3,5,4,4
2,1,4,2,2,2,3,1,1
3,1,4,3,1,1,3,2,3
4,1,1,2,4,4,6,5,4
5,1,5,3,3,2,4,2,3
6,1,2,4,4,3,4,3,3
7,1,2,3,2,3,4,4,5
19,1,3,1,2,2,2,2,1
20,1,6,2,2,2,4,3,2
22,1,4,1,2,2,3,1,1
25,1,2,2,1,1,4,1,2
26,1,5,2,2,2,4,2,2
29,1,4,3,2,1,3,2,2
30,1,4,4,4,2,1,2,1
32,1,4,1,3,2,3,2,3
33,1,4,2,3,2,3,2,3
37,1,4,2,2,2,3,3,3
run;


proc print data = made_from_log_file; run;

proc univariate data=made_from_log_file;
var convers; run;


data if_when; set made_from_log_file;
x = ranuni(0); /*assigns a value from 0 or 1 to each subject*/
run;

proc print data = if_when; var x; run;

Data ifdrop; set if_when;
if x > .2 then delete;
/*DROPS some WHOLE ROWS of data; I got ##--you will get more or less--why?*/
run;

Data ifkeep; set if_when;
if x > .2;
/*Alternative - KEEPS only observations specified*/
run;

Data if1; set if_when;
if x > .2 then convers = .;
/*recoding responses to missing for demonstration purposes*/
run;

proc print data = if1; var x convers; run;

proc means data = if_when; var convers; run; /*record this mean*/
proc means data = if1; var convers; run;

data if2; set if1;
if convers = . then convers = 3.32; 
/*example of 'mean substitution (bad idea) and recoding into the
same variable with an if then conditional statement*/
/*result--usually retains approximate value of the mean but biases the variance---How?*/
run;

proc means data = if2; var convers; run; 


data if_else; set if_when;
/*example of if--else*/
bobo = '              ';
if x > .2 and x < .35  then bobo = 'doll'; else bobo = 'not a doll';

bobo1 = .;
if x > .2 and x < .35  then bobo1 = 36; else bobo1 = 200;

green = .;
if x > .2 and x < .35  then green = 1; else if x > .45 then green = 2;
run;

proc print data = if_else; var x bobo bobo1 green; run;

/*Alternative symbols to use in 'if-then' conditional statements
=
NE
GT or >
GE
LT or <
LE
end of list*/



/*Example using two different variables in the if statement*/
data if_else; set if_when;
smile = .;
if happy LE 3 AND humor LE 3 then smile = 0;
else if happy GE 4 AND humor GE 4 then smile = 1;
run;

proc print data = if_else; var happy humor smile; run;


/*IN operator */
data if_else; set if_when;
glee = .;
if happy in (3,4,5) then glee = 17; else glee = 1;
run;

proc print data = if_else; var happy glee; run;


/*NOTIN operator */
data if_else; set if_when;
laugh = .;
if humor notin (2,4) then laugh = 17;
run;

proc print data = if_else; var humor laugh; run;


/* DATA MERGING */

/*In order to concatenate 2 data sets you need to be sure that both data sets are active
could be used to add participants to an existing data file*/
data filea;
input id var1;
cards;
1 10
2 20
3 30
4 40
5 50
;
run;

data fileb;
input id var1;
cards;
6 60
7 70
8 80
9 90
10 100
;
run;

data fileab; 
set filea fileb;
proc print data=fileab; run;


/*Merging 2 data sets
could be used to add variables to participants' data*/
data filec;
input id var1;
cards;
1 10
2 20
3 30
4 40
5 50
;
proc sort data=filec; by id; run;

data filed;
input id var2;
cards;
1 60
2 70
4 90
5 100
3 80
;
proc sort data=filed; by id; run;

data filecd; 
merge filec filed;
by id; 
run;
proc print data=filecd; run;


/* proc format and proc freq*/
proc format;
value warmcold
0 = ‘warm’
1 = ‘cold’
;
run;

proc freq data = if_when;
format warmcold warmcold.;
table warmcold;
run;
proc contents; run;