I have a following sample dataset and looking to check iteratively for each ID if the combinations on Var2 (eg: CC85 & CC86) or (cc54 & CC57) are present. Each ID can have one or more (than 5) observations. If those combinations are present then create a new variable called 'combo' and assign values like 'Hcc801' and  'Hcc501' respectively. No combo value for a single ID or if the conditions are not satisfied then the combo = 'NA'. Please note, there are many other combinations along with CC85 & CC86 or CC54 & CC57 that need to be incorporated.   data t1; input ID Var1 $ Var2 $ Rank; datalines; 1 N179 CC135 2 2 I214 CC85 1 2 A419 CC86 1 2 A500 CC100 3 3 A100 CC54 1 3 A101 CC57 ; run; Thank you in advance for your help! ... View more

Dataset looks like the following. There are 1000's of IDs and 100,000's of observations.  'Wanted Value' is the desired output using the variable "Value". The number of observations for ID is not constant, normally between 5 and 30. There could be intermittent missing values in between which need to be repopulated plus or minus the date difference, depending upon the moving forward vs backward. Missing values are labelled as ".". Thanks for your help!!     ID Date Value Date Diff Wanted Value 1 1/1/2000 . - 600-92-91-91 1 4/1/2000 . 91 600-92-91-91 1 7/1/2000 . 91 600-92-91 1 10/1/2000 . 92 600-92 1 1/1/2001 600 92 600 1 4/1/2001 600 90 600 1 7/1/2001 . 91 600+91 1 10/1/2001 730 92 730 2 1/1/2002 . 92 550-92 2 4/1/2002 550 90 550 2 7/1/2002 . 91 550+91 2 10/1/2002 600 92 600 2 1/1/2003 . 92 600+92 2 4/1/2003 . 90 600+92+90 2 7/1/2003 . 91 600+92+90+91 2 10/1/2003 . 92 600+92+90+91+92 2 1/1/2004 . 92 600+92+90+91+92+92 2 4/1/2004 . 91 600+92+90+91+92+92+91 ... View more