topic Re: Lag equivalent in Enterprise Guide in SAS Enterprise Guide

Lag equivalent in Enterprise Guide

altijani — Wed, 12 Sep 2018 14:55:43 GMT

Hi Community,

What is the function equivalent to Lag in enterprise guide? In other words, how can I create a new variable equals 1 if Var = Var from the previous observation?

Thanks!

Re: Lag equivalent in Enterprise Guide

RW9 — Wed, 12 Sep 2018 15:02:54 GMT

Lag function is a Base SAS function and is available for any version of SAS 9.2 upwards at least. So either your EG is very old, in which case upgrade, or something else is wrong.

Re: Lag equivalent in Enterprise Guide

Reeza — Wed, 12 Sep 2018 15:09:02 GMT

LAG doesn't work in SQL/Query Builder, it will work within a data step/SAS program. Alternatively, if you have SAS/ETS there may be timeseries procedures that allow you to create lagged variables via the GUI.

Re: Lag equivalent in Enterprise Guide

jebjur — Wed, 12 Sep 2018 15:19:18 GMT

The Enterprise Guide query builder uses PROC SQL under the covers, and the LAG function is not valid in PROC SQL. You would either need to use DATA step code in a program node, or there is a custom task you can use that will calculate the lag.

http://support.sas.com/documentation/onlinedoc/guide/customtasks/index.htm

Under the quick links for Eguide 4.3 and later, choose the link for the Example: Calculate Lag. Download the zip file to your PC, and save the CalculateLag.dll file to 'C:\Program Files\SASHome\SASEnterpriseGuide\7.1\Custom'.

When you open a new session of Eguide, under Tools - Add-in Manager, you should see the Compute Lag/Dif custom task.

You may need to unblock the DLL file on your PC as well, here is more information:

http://blogs.sas.com/content/sasdummy/2013/05/19/unblocking-custom-task-dlls/

Re: Lag equivalent in Enterprise Guide

TomKari — Wed, 12 Sep 2018 16:23:05 GMT

Just to expand a little on @jebjur's comment, it may seem weird that LAG doesn't work in PROC SQL.

It's important to know that SQL was originally intended as a "set" language, where the concept of processing one record after another doesn't exist. That's why the DATA step is so necessary for problems such as this. It's not a defect in SAS.

Tom