Looks like other community members beat me to the punch. I'd like to add it's a good idea to format/indent your code which makes it easier to debug data MSM_IDU; set testing_data ; format risk $10.; if x742 ne '1' then risk='UNKNOWN'; else if g124 in ('1', /* <-- NEED THIS COMMA HERE */ '3') then do; if g216a='1' then do; if g402='1' then risk='MSM/IDU'; else risk='MSM'; end; end ; /* <-- NEED THIS EXTRA END */ RUN; ... View more

Hi @ed_sas_member ,    Many thanks for your reply, I have a BSc in computer science so good to know that'd still be relevant. The book you recommended looks like a great place to start!   Best, ... View more

Correct. alpha=0.05 is the default so CIPCTLDF;   /* default is alpha=0.05 */ is equivalent to CIPCTLDF(alpha=0.05);   ... View more

INFORMAT is not intended to be used to DEFINE variables. It is instructions for how to READ the variable from TEXT. It will only define the variables as a side effect of being the first place you reference the variable in the data step, the same as side effect of using the variable in most other statements, like assignment statement.  Note that for most variable there is no need to attach any INFORMAT to the variable. SAS already knows how to read character strings and numbers from text.   The exception are things like DATE, TIME and DATETIME values where the text used to represent the value is totally different than the value that is stored in the variable. (see next point)   Do NOT use decimal part on an INFORMAT.  That is instructions to SAS for how to handle input strings without an explicit decimal place. So if you use an informat of 6.2 you are telling SAS to divide text that looks like an integer by 100.  So '123456' will be come the number 1,234.56 instead of the number 123,456.   I think you might have meant to use a positional variable list in your INPUT statement.  Something like: input cookie -- DSLV_etihad_guest_accounts ; Note that once the type of the variable has been defined there is no need to add an informat into the INPUT statement.  And if you do need to add an informat into the INPUT statement and you are reading from a delimited file then make sure to use the colon modifier before the informat to prevent SAS from reading the exact number of characters the informat spec says, even if it is longer (or shorter) than the next delimited value. ... View more

Hi @duckypooh    Two things: - you can use the IF/THEN statement outside a data step. That's why you get some errors in the log. TO avoid this, remove the "RUN;" just after the SET statement - another thing is that you need to define a sufficient length fo your categorical variables. Otherwise, they will be set to the length of the first value and it will cause unexpected results (truncation)   Best, data import; set mcdata.import; /*run; <---- remove run; */ if Preg_Group in ('Miscarriage', 'Normal'); length BMI_cat $10.; /* <---- set an appropriate length to avoid truncation */ if BMI > 28 then BMI_cat = 'BMI >28'; else if 19 <= BMI < 23 then BMI_cat = '19 <=BMI <23'; else BMI_cat = 'BMI Other'; length Age_cat $10.; /* <---- set an appropriate length to avoid truncation */ if Age <32 then Age_cat = 'Age <32'; else if Age >= 32 then Age_cat = 'Age >=32'; ln_Progestrone = log(Progestrone_nmol_L); run; ... View more

data have; infile datalines expandtabs; input Customer :$20. Year Widget_A Widget_B; datalines; CustomerA 2001 1 2 CustomerA 2002 3 4 CustomerB 2001 5 6 CustomerB 2002 7 8 ; proc sql noprint; select distinct catt('have(where=(year=',year,') rename=( Widget_A=Widget_A_',year,' Widget_B=Widget_B_',year,' ))') into : merge separated by ' ' from have; quit; data want; merge &merge; by Customer; drop year; run; ... View more

Proc sort data= have; by client_id date1 date2; Data want; Set have; By client_id date1 date2; Retain sum1 sum2; If first.date1 then sum1= value1; If last.date1 then do; Sum2 = value2; Output; End; Run; ... View more