Take the 2-minute tour ×
Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It's 100% free, no registration required.

I have the following column:



Where dots indicate that there was no data available at that time.

Now I want to compute the difference in growth between the current case and the last available case. Thus, in the second last case this should become: 0.12 - 0.42 = -0.30.

Lag(growth) takes the last case, and returns nothing if it is not defined. So how do I get the last available case?


This seems to work:

define !lastAvailable(target !Tokens(1) /level !Tokens(1) /Varys !CMDEND)
   compute id= $casenum.
   dataset copy tmpset2.
   select if not miss(!target).
   compute diff= !target - lag(!target, !level).
   match files /file= * /file= tmpset2 /by id.
   dataset close tmpset2.

!lastAvailable target=growth level=3
share|improve this question
What are you really trying to do? Perhaps RMV (Transform>Replace Missing Values) would be useful here. –  JKP Jan 8 '12 at 17:01
I'm afraid not, these values cannot be computed because they should not exist. In real life, there are no values at these intervals either. –  Tom Jan 8 '12 at 17:02
add comment

1 Answer

up vote 4 down vote accepted
if not miss(lag(growth)) #x= lag(growth).
comp diff= growth-#x.


Addition. The "general" solution that you ask for in the comment and that should allow for arbitrary lag seems hard to achieve in the context of presence of missing values not following a specific pattern. So, the straightforward way would be to get rid of missing values temporarily; without them, the task becomes trivial.

comp id= $casenum. /*Create case identificator
dataset copy set2. /*Take a copy of the dataset
select if not miss(growth). /*Delete missing cases from original dataset
comp diff= growth-lag(growth,3). /*And compute difference with the lagged case (in this example, lag 3)
match files /file= * /file= set2 /by id. /*Insert back missing cases from set2
dataset close set2. /*Delete set2
share|improve this answer
What language is this? It does not seem to run in SPSS. Also, if lag is missing, how will it take the last available value? Did you understand my question? –  Tom Jan 7 '12 at 14:34
This is clean SPSS syntax. I checked it on SPSS 20 and the result is what I perceive you wanted in your question. Now I show the resultunt column in my answer. –  ttnphns Jan 7 '12 at 15:05
You are correct. I should have mentioned that I also need to be able to get the second last item, third last item, etc. I figured there was a general method for this. I understand why your logic works for the last available item, but can it be applied for the second last available item? –  Tom Jan 7 '12 at 15:41
So, you want to pick, say, the 2nd (not the 1st) value above the current value and, if that value is missing, then pick the closest above it. Right? Then change both lag(growth) in the first command to lag(growth,2). Will that suit you now? –  ttnphns Jan 7 '12 at 16:21
No, I want to get the second last available value. So, in my example it should return 0.12 - 0.23 = -0.11 for the second last row (the first last available value is 0.42, the second last available value 0.23 and the sixth last available value is 0.5). –  Tom Jan 7 '12 at 16:26
show 7 more comments

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.