Getting started with the community

How to get fast, helpful answers

by Community Manager ‎09-17-2015 04:42 PM - edited ‎04-13-2018 05:35 PM (5,330 Views)

In the SAS Support Communities, you want correct answers. Fast. Well-crafted questions attract fast answers -- but you better not post them as new topics if you want speedy help.

 

The worst thing you can do...guaranteed to bring your progress to a screeching halt...is to post a new question on an old discussion thread. Nobody's looking for new questions on old threads!

 

Before you post, search to see if someone already answered your question in the community. If not, choose the support communities board that best matches your topic and ask your question there.

 

3 steps to compose a great question

 

1. Write a descriptive subject line. Be specific. "How Do I Extract a Submatrix?” is better than “Newbie Question.”

 

2. Use simple language. You might be an expert in non-parametric statistics, but other readers may not be. Explain your problem simply and in context so that a reasonably smart person can understand.

 

3. A good question includes code and example data.  Here's an example of how to create a data step of your data. It's quite helpful! Even better, include what your output should look like. Remember not to include any confidential data in your posts or anybody’s personal information.


Now, let’s compare two ways to ask the same question. First example:

 

question_1.PNG

 

Have you ever posted a question like this? How’d that work out for you? Looks like somebody procrastinated on an assignment! Seriously, though, this question’s too vague. It forces fellow community members to follow up and ask more questions to fill in the gaps. Take a look at the answer this question attracts:

 

Answer_1.PNG

 

It’s not an answer at all, but a series of clarifying questions. If enough detail had been provided the first time, the correct answer would’ve come faster.


Here’s a better way to ask the question:

 

Question_2.PNG
This question is ideal because it mentions everything relevant. Notice that it includes code to illustrate the problem. This question attracts a complete, helpful answer:

 

Answer_2.PNG

 

 Here are a couple more tips:


• Check “like” beneath answers and community articles you find helpful. This makes the best content rise to the top of community searches where other users easily find them…and rewards those who help you.

 

• Ask questions even if English isn’t your first language. Sometimes posting code explains your problem better than a thousand words. If context is needed, post in your native language. Fellow users or the community managers will figure it out.

 

If you’d like more detail, here's an excellent SAS Global Forum paper on this topic.


SAS Support Communities exist to help you increase your SAS knowledge. Help us help you by asking the best question you can!

 

The following is for you visually oriented people, if your workplaces allow you to watch video: 

 

 

 

Comments
by Respected Advisor
on ‎09-17-2015 07:10 PM

Also, if there is an error message, don't say "I get an error".

 

Show us the relevant portions of the SASLOG, the code that is in error and the error itself.

by Community Manager
‎09-18-2015 02:01 PM - edited ‎09-18-2015 02:02 PM

Good point, @PaigeMiller!

by Occasional Contributor ablo
on ‎12-21-2015 07:54 PM

How do I use the sample data sets for SAS base programmer? I dont see any assignement that work with data. I am just wondering if the data are just provided like that?

by Community Manager
on ‎12-21-2015 08:53 PM
Hi @ablo, are you just beginning to learn SAS? If so, post your inquiry in the Analytics U Community, where lots of helpful experts can see it and respond.
by Occasional Contributor DorotheaBeck
on ‎04-04-2017 07:44 AM

Hi,

 

I am a SAS programmer / data manager at Heidelberg National Center for Tumor Diseases (nonprofit organization), and I want to use some SAS code that was posted in a community forum.

 

How do I find out about copyright and the correct way to acknowledge the author?

 

Thanks in Advance,

 

DBeck

 

 

by Community Manager
on ‎04-04-2017 02:31 PM

Hi @DorotheaBeck,

 

If the code was posted on this community, I suggest you send a private message to the author of the code/article and start there. 

 

Good luck and best wishes,

Shelley

by Occasional Contributor DorotheaBeck
on ‎04-05-2017 02:25 AM
Hi Shelley,

thanks for the answer. I had thought of doing that but did not dare at first because it said everywhere, do not write private messages...

Regards,

Dorothea
by Respected Advisor
on ‎04-05-2017 08:28 AM

The advice to not write private messages is that you shouldn't ask for help solving a problem in a private message, these ought to be public in the Communities so anyone can read the answers and anyone can join the conversation.

 

If you have a question about permission to use code, this is something that belongs in a private message.

by Occasional Contributor DorotheaBeck
on ‎04-06-2017 02:00 AM
Thanks for the info. I had somehow suspected that but was not sure...
by Super User
on ‎04-02-2018 07:16 AM

It would be great if you could include a link to the post that shows how to convert a dataset to a data step for posting example data.

by Community Manager
on ‎04-02-2018 09:28 AM

Great suggestion, @KurtBremser. Since @BeverlyBrown is out this week, I've edited the article and included that link.

by Community Manager
on ‎04-07-2018 06:39 PM
by PROC Star
on ‎04-07-2018 08:13 PM

While one would think it would be obvious, obviously it isn't! I'd begin the article with "Never, ever post a new question to an existing or someone else's thread .. and, if it's your thread, create a new thread if you've already marked the question as solved."

 

Art, CEO, AnalystFinder.com

 

by Community Manager
on ‎04-13-2018 04:48 PM

Good point, @art297. I'll work that in!

by Occasional Contributor IliyaGB
on ‎05-04-2018 02:36 PM

%macro statut_client(data,clientid);
proc print data=&data;
var clientid duration purpose savings_status class;
where clientid=&clientid;
%if class="good" %then %do;
proc print data=&data;
var clientid duration purpose savings_status class;
where clientid=&clientid;
%put "bon client";
%end;
%else %if class="bad" %then %do;
%put "mauvais client";
proc print data=&data;
%end;
run;
%mend;
%statut_client(data_credit,1215693);

 

Trying to run this macro but it's not working properly. I need clientid duration purpose and savings_status when class=good and I want a message to appear "bad client" when class=bad.

Can someone help me, pls?

by Community Manager
on ‎05-04-2018 02:49 PM

Hi@IliyaGB, you'll get an answer faster if you post your question on one of our discussion boards. Go to communities.sas.com and choose from the list of boards under "Find a Community." Click the + under SAS Programming to see a list of forums. All the best to you!

by Super User
on ‎05-16-2018 03:14 AM

Why do we want example data in a data step?

 

Compare the three methods for posting example data most commonly found:

 

  • taking a screenshot. Basically the WORST option by a mile. It consumes the poster's time by having to upload a picture and inserting it in the post, and forces everyone else to type letter-by-letter off the screen. Actually considered rude by many long-time members here.
  • doing a copy-paste from a viewtable window or similar. Much better, as it is quick to post, and data can be copy-pasted into program code, but if posted in the main window (not one of the code windows), content can be changed. And the delimiter is not always obvious. And people still need to write a wrapper program around it to get the dataset, while having to guess for column attributes like lengths and formats.
  • writing a data step with datalines, or using the macro mentioned, and posting the code in a code window. Everybody can recreate exactly the same dataset with a copy/paste and submit. While this takes some effort from the poster, it makes solving issues a breeze in most cases, as testing the suggested code is just a mouse-click away.
    And for a newbie, writing such a datalines step can be a great first learning experience in itself. It was (and sometimes still is) for me.