I am not exactly sure how to describe this, but I have a panel dataset of loans that gives the loan age by month. Occasionally there are gaps, and I want to record in a new variable the total number of gaps. For instance, see the data below:     I would like to save, by loan_id, the different gaps in x_loan_age. So I would like to save that there is a 38 month starting in 200508 (YYYYMM variable), a 9 month gap starting in 200801, and a 15 month gap starting in 200810.   Ideally I would create variables called "GAP1", "GAP1START", "GAP2", "GAP2START", "GAP3", and "GAP3START".   The output would then be: 38, 200508, 9, 200801, 15, 200810.   There are tons of loans in this dataset, so I want to do this by each loan_id.   Basically, this is a dataset of when people didn't pay their loans and the gaps are when the person actually paid. I am trying to discern when they were paying among these missed payments.    If this can't be handled by sas, alternatively I would just like to count the cumulative "gap" space. So for the above loan, returning a variable called "GAPS" that = 38+9+15.    Attached is a SAS dataset of the data I pasted. Thanks!     ... View more

Hello, I have two datasets that look like the following:   Dataset 1: sbdv num loc 150 0.876 1 160 0.993 1 171 0.937 1   Dataset 2: sbdv num loc 120 0.987 0 121 0.965 0 123 0.945 0 125 0.912 0   I want to create a new dataset that contains all of dataset 1 plus a match of observations in dataset 2 of the observations with the minimum difference in 'num' between each obs in dataset 1 and all the obs in dataset 2.   So for sbdv = 150, I want to calculate the difference between its 'num' (0.876) and sbdv 120, 121, 123, and 125's respective 'num' values (0.987, 0.965, 0.945, 0.912). Then I want to save each 'match' to a new dataset containing all of dataset 1 and the matches for each observation.   So for this example, the new dataset would be: sbdv num loc 150 0.876 1 160 0.993 1 171 0.937 1 125 0.912 0 120 0.987 0 123 0.945 0   I am guessing there is a straight forward solution, but I haven't been able to find anything that explicitly will do this. ... View more