i have a dataset with the following variables: bmicat(ordinal with 3 categories), then age, sex, time, smoking. This is a longitudinal and incomplete dataset (time = 0,1,2,3,4,5,6,7,8,9), sex(0,1), age is continuous) and smoking(0,1). Now i want to do multiple imputation under missing at random then fit a generalised linear mixed model (proc glimmix data=Tmp1.Bmilda_bmicat_baseline method=QUAD (qpoints=20); title 'PROC GLIMMIX analysis, ordinal, AGQ (PQL, REML)'; class time id sex baseline_smoking; nloptions maxiter=10000; model bmicat = timecont sex fage baseline_smoking timecont*sex timecont*fage timecont*baseline_smoking/ dist=multinomial link=cumlogit solution; random intercept timecont/ subject=id solution g; output out=random pred=p ; run; or the nlmixed version). I keep getting errors I can't fix. The 3 steps i want to follow are: 1. proc mi (multiple imputation) 2. PROC glimmix or nlmix 3. proc mianalyse. Please provide SAS code that works, you can test it on sample data. Please don't give me references, I need working code