topic Re: SAS Code Errors in SAS Programming

SAS Code Errors

BrahmanandaRao — Mon, 04 Apr 2022 09:55:34 GMT

Hi Guys

An interview panel asked me what do you do if your code has many errors

you identified after send to the user how to resolve the issue

Re: SAS Code Errors

PaigeMiller — Mon, 04 Apr 2022 10:18:37 GMT

Maxim 2 — Read the log

Of course, this part of the question:

you identified after send to the user

if I am understanding the question, is not really a SAS question, it is a question about how you behave after you discover you have made a mistake, what do you do, what do you say? Please give an answer, and those of us here (such as me) who have been in that exact situation can critique your answer.

Re: SAS Code Errors

Ksharp — Mon, 04 Apr 2022 11:38:07 GMT

The first ERROR is very important .

Re: SAS Code Errors

BrahmanandaRao — Mon, 04 Apr 2022 12:28:12 GMT

First Error means

how to convey to user our mistake

Re: SAS Code Errors

PaigeMiller — Mon, 04 Apr 2022 12:50:05 GMT

Hello, @BrahmanandaRao

A word to the wise: if you are on an interview, this type of brevity (and frankly this type of unclear communication) is not a good idea. It is also not a good idea here in the SAS Communities, as you will get faster and better answers from questions/comments which are clear and complete, rather than from questions/comments that are unclear and incomplete.

Re: SAS Code Errors

Patrick — Mon, 04 Apr 2022 13:34:18 GMT

@BrahmanandaRao wrote:

First Error means

how to convey to user our mistake

My 5 cents: Everybody makes mistakes. Worst thing you could do is try to hide it. Be pro-active, contact the user, apologize, provide a solution. ...and learn from your mistake and improve your testing/quality checking.

And if the mistake is serious and could have commercial consequences: First inform your manager so the right course of action can be taken.

Re: SAS Code Errors

Kurt_Bremser — Mon, 04 Apr 2022 14:17:37 GMT

Step 1: try to minimize the consequences. This means that all people that might start to work with incorrect data need to be notified ASAP. If in doubt, the broader notification is better; one may want to employ the helpdesk for sending out the notification. Been there, done that.

Step 2: face up to the issue. The only one who never commits mistakes is one who does no work at all. Been there, done that.

Step 3: prevent a repetition of the issue. See why a code was able to create incorrect results without throwing a recognizable error (ERRORs/WARNINGs are caught by the scheduler which then suspends further processing, and notifies at least the data center standby people, who then notify you). A large part of my work was committed to this, which in turn freed me from having to constantly inspect logs and datasets in search of possible problems.

As soon as you have correctly identified the source of the problem, corrected it, and can present a viable method you implemented to prevent a repetition, you'll be good with your bosses and customers.