BookmarkSubscribeRSS Feed

Hello - 

 

I would like to make the suggestion that once you update a name of a column for any multitude of reasons you should have the ability to auto-update that variable name in all the following datasets.  If you need to change a variable name in an earlier dataset the datasets that follow will break if you run the process flow due to the variable 'not existing'.  It would be awesome to not have to manually update the variable name in the rest of the datasets, because that could be very tedious and a lot of work.  I know that you can simply name the last set or output with the new name, but I'd like to be able to run with the correct variable name through-out the whole flow.

 

Thanks 

 

Andrew 

3 Comments
Haikuo
Onyx | Level 15

I suppose you are talking about the query builder behavior, and the answer is a big NO. If you see the code generated by the query builder, all the requested columns are selected explicitly, even the case when you need all of the columns. There are underlined reasons for SAS choose to do this (select * vs select columns), the advantage of choosing 'select colunms' over 'select *' becomes obvious when you deal with RDBMS data, which is a common task for SAS. Please feel free to google the reason if you are interested.

Now get back to your problems, the only way to get around is to code it such as (select *) all along in your project after you rename your variable. And I am also puzzled by your statement, "but I'd like to be able to run with the correct variable name through-out the whole flow." Why?

 

PhilC
Rhodochrosite | Level 12

I think this is a horrible torturous thing to ask the developers to do...   But I would SO use it.  To my defense, this is the only time I've ever supported torture.

ChrisHemedinger
Community Manager
Status changed to: Not Planned

SAS Data Integration Studio (and related SAS Data Management tools) are much better suited to support the discipline of following the chain of changes and impact analysis.  Table and column metadata is a baked-in component of those tools.  However, SAS Enterprise Guide doesn't track the data and column use to that level -- and there aren't near-term plans to add it.  For folks who need this level of governance, we usually recommend the data management solutions.

 

(I appreciate @PhilC's concession that this might be a difficult thing for R&D to implement.)