I have a dataset where it appears some ht and wt values were in put in the wrong column (e.g., ht in wt and vice versa).    My gorilla code (see: use brute force) below worked, but I'm wondering if I could do this more succinctly, i.e., swap ht and wt values for certain observations meeting conditions specified by my IF statements.   I also recognize I could combine the two datasteps below into one, and subsequently did, but wanted to be able to check my code was doing what I wanted in between. This is the same reason I didn't drop my htnew2/wtnew2 variables.  data htwt; set wtnew; htnew2=0; /*added because otherwise this variable was uninitialized*/ wtnew2=0; /*added because otherwise this variable was uninitialized*/ IF 0<htnew<150 and wtnew>110 then htnew2=wtnew; IF 0<htnew<150 and wtnew>110 then wtnew2=htnew; run; /*I then checked these variables with proc freq; also ran proc print of those observations*/ data htwtcorrect; set htwt; IF wtnew2>0 and htnew2>0 then htnew=htnew2; IF wtnew2>0 and htnew2>0 then wtnew=wtnew2; run;   ... View more

I have character variables that have numeric values, such as height and weight. To recode these as numeric, I made a new variable using the 'input' such as below.  weightnew = input(weightold, 8.); Functionally this works, but it also produces an error wherever there is a missing value for weightold.  NOTE: Invalid argument to function INPUT at line 10516 column 10. WARNING: Limit set by ERRORS= option reached. Further errors of this type will not be printed. This ultimately doesn't appear to matter as each of those missing character values becomes recoded as a missing numeric value, but the sight of all those errors in the log makes me worry something is wrong.    Is there a better way to change weightold (character) to weightnew (numeric) than using the input statement?   Secondly, should I be bothered by the error log looking the way it does, or if it is doing what I want should I just ignore?    Thanks.    ... View more

I have a data set where a "count" variable means there are X numbers of individuals with the preceding characteristics in that observation. See below as an example. In the first observation, 3 means there are three 24 year-old males who scored 100.    data sample; input age score sex$ count; datalines; 24 100 M 3 25 95 F 2 25 97 M 2 20 86 M 3 30 100 F 2 ; RUN; Let's say I wanted to do a proc freq of sex * score.   PROC FREQ data=sample; TABLES sex*score; run; Is there a way to call out the actual number of each sex achieving each score by using the "count" value for each observation? My code above counts only one sex per row, obviously.  ... View more

I am creating indicator variables for character variables with a dataset imported from excel. In this dataset there are multiple misspellings, extra spaces, etc. so I copied and pasted each value of the variable from a PROC FREQ to get something like the below:      IF oldvar='oldvaluex' OR IF oldvar='olddvaluex' OR IF oldvar='oldvaluex ' ...etc. then newvar=0; ELSE if oldvar='oldvaluey' OR IF oldvar='old valuey' OR IF oldvar=' oldvaluey '... then newvar=1; ELSE newvar='.';     In comparing oldvar with newvar using PROC FREQ I noticed some variations of 'oldvar' were being coded as missing. In searching through these forums I found that it was likely a spacing problem and by using oldvar=compress(oldvar,,'C') prior to creating my indicator variable, these values were now appropriately coded to newvar.    However, I am now working on another variable which is still producing missing values after using oldvar2=compress (oldvar2,,'C') and also attempting oldvar2= Compress(oldvar2,'0D0A'x).    Any ideas on what else I could try?  ... View more

I am using PROC SQL to try to bring some additional variables from each data1 (var1-5) observation into a new dataset when there is a match with a data2 observation across var1-3.   (For my real world data, data2 is the entire sample of subjects and data1 is a small percentage of the sample who had some additional measures. There is no common ID to join by, so I am trying to join where 6 demographic variables are the same in both datasets.)   Goals:  1. Join on a one-for-one basis (i.e. one data1 observation is joined with one data2 observation, and var4-5 from data1 are brought into new dataset). 2. Retain all data2 observations where there was no match, with missing variables for var4-5. 3. Know when/which data1 observations did not have a match, and were thus not included in the new data set.   Problems:  1. My current code is bringing data1 var4-5 to ALL data2 observations where var1-3 match. Is a left join the wrong operation altogether, or are my x=y conditions able to be modified to restrict to one match?  2. This works with my current code.  3. Haven't figured out a clean way to identify this, so not included in my code.    Sample code below (SAS 9.4):  data data1; input var1 var2 $ var3 var4 var5; cards; 1 A 1 1 1 1 B 2 2 2 1 C 3 3 3 1 D 4 4 4 ; run;   data data2; input var1 var2 $ var3; cards; 1 A 1 1 A 0 1 B 2 1 B 2 1 C 3 1 C 4 ; run;   proc sql noprint; create table new as select * from data2 x left join data1 y on x.var1 = y.var1 and x.var2 = y.var2 and x.var3 = y.var3; quit;   New Dataset: var1 var2 var3 var4 var5 1 A 0 . .  1 A 1 1 1 1 B 2 2 2 1 B 2 2 2  *This is what I am trying to eliminate. If no other matches in data1 this line should be "1 B 2 . ."* 1 C 3 3 3  1 C 4 . .    *I would like to be able to see that observation 4 from data1 was not included in the new dataset (1 D 4 4 4).* ... View more