Many thanks for all your responses, been a bit busy so only had chance to respond.   So, I managed to solve my issue with the following set statements:   data want; set have; set have(firstobs=2 keep=newid aeid linkid rename=(newid=x1 aeid=y1 linkid=z1)) set have(obs=1 drop=_all_); if x1=aeid then linkid=z1; run;   So for the rash maculo papular events I kept my retain, and then used the above to correct the population of the LINKID when NEWID doesn't run chronologically.   Happy to consider this closed now 🙂   Thanks alot. ... View more

Thank you for the above response 🙂   However, I would expect it to look like follows (I have also added how it would look like the for EVENT mentioned at the beginning of this thread):   ID NEWID EVENT AEID LINKID ------------------------------------------------------------------------ 101 7 Rash_maculo-papular 5 5 101 12 Rash_maculo-papular 7 5 101 15 Rash_maculo-papular 12 5 101 . Rash_maculo-papular 15 5 101 19 Rash_maculo-papular 18 18 101 . Rash_maculo-papular 19 18 102 . Dyspnoea 1 2 102 1 Dyspnoea 2 2   So for ID 101 when NEWID = 7 and EVENT = Rash_maculo-papular this is classed as the first occurence of the EVENT within the linked EVENTS - so LINKID is populated with 5 already.   And then for ID 101 when NEWID = 19 and EVENT = Rash_maculo-papular this is classed as the first occurence of the EVENT within the linked EVENTS - so LINKID is populated with 18 already.   So then when NEWID = AEID I'd like LINKID to be populated with the LINKID from the first occurence (as shown above)   This was done fine in the code I was using, when using a retain function. However, the second example I have given causes problems as I cannot use a retain function on that.   For ID 102 when NEWID = 1 and EVENT = DYSPNOEA this is classed as the first occurence of the EVENT, so LINKID is populated with 2 already for that observation. So I'm wanting the record above it, where the NEWID = AEID to be populated with the LINKID from the first occurence (which would be 2).   So I'm hopefully looking for a solution with the outcome of the table above.   I really appreciate the help so far!, it's just a really awkward one to try to explain, so hoping the above can clear it up abit 🙂     ... View more

I will have to attach an excel file, due to working off a company laptop the code that has been provided to convert to the dataset to SAS code will not work.   The attached excel file, presents all the records in the dataset. Which you should be able to import into SAS.   So... when NEWID is present (for the very first ADVERSE EVENT of a set of linked adverse events) the LINKID will equal the AEID. The above has already been created in the dataset.   Then within the same linked adverse events, when NEWID is equal to the AEID, the LINKID needs to equal the LINKID of the first ADVERSE EVENT. ... View more

SUBJECT ID NEWID ADVERSE EVENT AEID LINKID 101 7 Rash maculo-papular 5 5 101 12 Rash maculo-papular 7   101 15 Rash maculo-papular 12   101   Rash maculo-papular 15   101 19 Rash maculo-papular 18 18 101   Rash maculo-papular 19     So I've added a further example - within a certain ADVERSE EVENT, when the ADVERSE EVENT is linked the NEWID will always match up to AEID, generally it can fall on any record.    As you can see above the LINKID which I want to display on the linked ADVERSE EVENT is the first value of AEID within the linked ADVERSE EVENT, so I was just using the retain function to populate the missing LINKID's   However, as per the example in the first post, the retain function would be retaining a value from a previous (different ADVERSE EVENT) which is giving me incorrect results. ... View more