BookmarkSubscribeRSS Feed
knveraraju91
Barite | Level 11

Dear,

 

I need help in my code in data step three.

 

First I sorted data and ran a data step to get last record for 'id test date'.

 

Then I need find if 'test' in lab1 and lab2 and at least one obs where value= 'neg' then want='neg';

else if 'test' in lab1 and lab2 and at least one obs has the vaiable 'value'  other than neg or blank then want='pos';

if 'test' in lab1 and lab2 and both values of variable 'value' are blank then want='un';

 

Thank you

 

output needed;

id test value date           want
a lab1 neg 2015-10-10 neg a lab2 neg 2015-10-10 neg b lab1 neg 2015-10-01 neg b lab2 2015-10-01 neg c lab1 pos 2015-10-01 pos c lab2 2015-10-01 pos d lab1 eqi 2015-10-01 pos d lab2 2015-10-01 pos e lab1 2015-10-01 un e lab2 2015-10-01 un

 

input

data one;
input id $1 test $3-6 value $8-10 date$11-21;
datalines;
a lab1 neg 2015-10-01
a lab2 neg 2015-10-01
a lab1 neg 2015-10-10
a lab2 neg 2015-10-10
b lab1 neg 2015-10-01
b lab2     2015-10-01       
c lab1 pos 2015-10-01
c lab2     2015-10-01
d lab1 eqi 2015-10-01
d lab2     2015-10-01
e lab1     2015-10-01
e lab2     2015-10-01
;
proc sort data=one;
by id test date;
run;

data two;
set one;
by id test date;
if last.test;
run;
proc sort data=two;
by id date test;
run;

data three;
set two;
if test and lag(test) in ('lab1' 'lab2') and value and lag(value) in (('neg' 'neg') or ('neg' '')) then want='neg';
else if test and lag(test) in ('lab1' 'lab2') and value and lag(value) not in (('neg' 'neg') or ('' '')) then want='pos';
else if test and lag(test) in ('lab1' 'lab2') and value and lag(value) not in (('' '')) then want='un';
run;

 

3 REPLIES 3
Shmuel
Garnet | Level 18

It is not recommended to use lag function in a conditional statement.

Try next code (the 3rd step only):

data three;
 set two;
  by id;
     retain lag_test lag_value;
     lag_test = lag(test);
     lag_value = lag(value);

     if test ne lag_test then do; 
        if (value='neg' or lag_value='neg') then want='neg'; else
        if (value=' ' and lag_value=' ')  then want='un'; else want='pos';
     end;
     drop lag_test lag_value;
run;

 

art297
Opal | Level 21

Not sure if I follow your rules, but the following achieves your desired result for your example:

data one;
  informat date yymmdd10.;
  input id $ test $ value $ date;
  format date date9.;
  datalines;
a lab1 neg 2015-10-01
a lab2 neg 2015-10-01
a lab1 neg 2015-10-10
a lab2 neg 2015-10-10
b lab1 neg 2015-10-01
b lab2 .   2015-10-01       
c lab1 pos 2015-10-01
c lab2 .   2015-10-01
d lab1 eqi 2015-10-01
d lab2 .   2015-10-01
e lab1 .   2015-10-01
e lab2 .   2015-10-01
;

data two;
  set one (where=(test in ('lab1','lab2')));
  if value eq 'eqi' then wanta='pos';
  else if missing(value) then wanta='un';
  else wanta=value;
run;

proc sort data=two;
  by id descending test wanta descending date;
run;

proc sort data=two nodupkey;
  by id descending test wanta;
run;

data want;
  do until (last.id);
    set two;
    by id;
    if wanta eq 'pos' then pos=1;
    else if wanta eq 'neg' then neg=1;
    else if wanta eq 'un' then un=1;
  end;
  do until (last.id);
    set one;
    by id;
    if neg then do;
      if not pos and not un then do;
        if first.id then output;
      end;
      else do;
        wanta='neg';
        output;
      end;
    end;
    else if pos then do;
      wanta='pos';
      output;
    end;
    else if not pos and not neg then output;
  end;
run;

Art, CEO, AnalystFinder.com

 

Kurt_Bremser
Super User

Every call of the lag() function retrieves a value from the FIFO chain and feeds a new value into it, so you must call lag() only once per data step iteration for a given variable.

And

if test and

means that test is converted to numeric and used as a boolean value, which is not what you intended IMO.

Maxim 2: Read the log! It will alert you to such unwanted effects.

sas-innovate-2024.png

Don't miss out on SAS Innovate - Register now for the FREE Livestream!

Can't make it to Vegas? No problem! Watch our general sessions LIVE or on-demand starting April 17th. Hear from SAS execs, best-selling author Adam Grant, Hot Ones host Sean Evans, top tech journalist Kara Swisher, AI expert Cassie Kozyrkov, and the mind-blowing dance crew iLuminate! Plus, get access to over 20 breakout sessions.

 

Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 3 replies
  • 762 views
  • 3 likes
  • 4 in conversation