Doc_Duke Tracker

Re: Chi-Square WARNING

Doc_Duke — Wed, 09 Mar 2022 21:06:23 GMT

It may be that a categorical test is not appropriate to these data. PROC NPAR1WAY seems more appropriate for testing the difference between waves on age. Though it will work with grouped ages, it would be even better if you have discrete ages.

Re: Exposure-adjusted event rate

Doc_Duke — Tue, 13 Apr 2021 00:33:39 GMT

Just do a google search and you'll get the answer. Lex Jansen has a nice article on it.

Re: Time to event analysis- ADTTE

Doc_Duke — Wed, 17 Feb 2021 21:49:52 GMT

Because the ADTTE is CDISC standard. See

https://www.cdisc.org/standards/foundational/adam/adam-basic-data-structure-bds-time-event-tte-analyses-v1-0

Standards can be changed, but it is a bit of a drawn out process.

Re: How can I Find Code that opens a given dataset?

Doc_Duke — Wed, 17 Feb 2021 21:37:40 GMT

This may not be viable, depending on how many EG projects you have, but you could run the project and store the log(s). Then you can just use text search tools to find the datasets referenced in the logs.

Re: How to extract date from start date (description part) and create a variable in the dataset

Doc_Duke — Sun, 15 Nov 2020 16:09:46 GMT

I get access denied on the google sheet.

Re: ERROR: No valid observations are found.

Doc_Duke — Fri, 18 Sep 2020 20:36:27 GMT

"ERROR: No valid observations are found." means that every observation is has a missing value for one or more of the predictor or outcome variables. Since it occurs for the continuous variables only, that's the place to focus.

Your dummy variable regression is likely overparameterized. Unless you have a lot of missing gender values, you don't need indicators for both male and female. Ditto gas and race.

You may want to try PROC GLM unless you ultimately want to use some of the stepwise parameters of PROC REG.

Re: setting VALIDVARNAME as the default for the duration of the session

Doc_Duke — Mon, 08 Jun 2020 18:10:05 GMT

Try putting it in the autoexec.sas file. I'm not currently using SAS Studio, so it's more a guess based on other SAS products.

Re: SAS Enterprise Guide Project Size Increasing with no changes

Doc_Duke — Mon, 13 Apr 2020 14:36:18 GMT

There are a few things that might be causing this:

-- Are you using the same version of EGuide? Your saving might convert to a newer storage.

-- Do you and your colleague have the same options set? I'm not sure which ones might impact storage.

-- Is your computer's storage blocksize different from your colleagues? I wouldn't expect that big a shift, but it might.

-- Are you using the same version of Windows? That might also change storage settings.

Re: How to calculate confidence interval for crude rate by using age group data by years

Doc_Duke — Mon, 13 Apr 2020 12:58:12 GMT

This may be the answer.

https://communities.sas.com/t5/Statistical-Procedures/Confidence-intervals-for-rates-using-Poisson-distribution/td-p/25521

I used this google search to find it

poisson confidence limit site:sas.com

Re: Why will my proc lifetest code not work?

Doc_Duke — Thu, 26 Mar 2020 19:01:16 GMT

PROC LIFETEST uses lots of memory. Look for 'memory requirements' in the documentation

https://support.sas.com/documentation/onlinedoc/stat/131/lifetest.pdf

Re: DateTime Calculations

Doc_Duke — Fri, 24 Jan 2020 20:25:40 GMT

To add to Reeza's comments, I'd suggest reading

https://documentation.sas.com/?docsetId=lrcon&docsetTarget=n0q9ylcaccjgjrn19hvqnd9cte8p.htm&docsetVersion=9.4&locale=en

for thorough coverage of the concepts.

Re: Creating a Variable

Doc_Duke — Tue, 10 Dec 2019 23:02:24 GMT

It appears that RET is a variable and you are using it as a function.

Re: Insufficient Space in file and i/o processing did not complete

Doc_Duke — Wed, 04 Dec 2019 00:34:53 GMT

In addition to Kurt's comments, you may be creating a number of intermediate files that are no longer needed, but are still taking up space on the disk. Critically look at your code and perhaps add some clean-up steps.

Re: Different means between PROC MEANS and EXCEL

Doc_Duke — Sun, 01 Sep 2019 19:59:40 GMT

Those difference look like round-off errors. although both SAS and Excel use double precision FLOATING POINT arithmetic, you can get round-off errors when the numerator N is large.

Re: side by side BOXPLOT

Doc_Duke — Sun, 01 Sep 2019 19:49:45 GMT

Lacona,

To get the boxplot you are looking for, you need to transpose the data first to get 3 columns:

Name, IncomeClass (with values Income1, Income2, and Income3), and the value of the income.

The elegant way to do it is to use PROC TRANSPOSE. However, with this little data, you can just retype the data.

Re: Heckman model.

Doc_Duke — Tue, 27 Aug 2019 01:37:06 GMT

This discussion might help

https://communities.sas.com/message/227949#227949

Re: EDT(Electronic data Transfer)

Doc_Duke — Sun, 23 Jun 2019 20:01:44 GMT

Reeza's comment to talk to a colleague in the area is spot on. The procedures and skills needed vary depending on the goal of the EDT. It could be that you need to do a lot of ODS formatting for report generation back to your clients or that you need ODBC or Oracle connectivity to load a database.

Re: Disk full OUT OF RESOURCES question

Doc_Duke — Thu, 20 Jun 2019 19:08:13 GMT

You probably need to look at some coding efficiency activities. For instance, using TAGSORT takes lot less disk space than a regular sort. You may also want to explicitly delete work files and, when possible, use the KEEP option so the procedures minimize the data they have to carry.

Re: Problem with tab delimited data file

Doc_Duke — Thu, 20 Jun 2019 18:51:54 GMT

CH33,

Another approach might be to use an editor that show the actual ASCII codes (e.g. UltraEdit). Then you can truly see the differences in the files.

Re: EDT(Electronic data Transfer)

Doc_Duke — Thu, 20 Jun 2019 18:45:35 GMT

In the clinical trials arena, EDT generally refers to transmitting data without a lot of human oversight. It can range from a nightly run program that takes the days analyses from the lab and packages it as reports to the sites to a more real time activity that monitors for incoming data, processes it, and send the results out.

One of the big differences I see is that the program(s) need to have a lot more error logic and alerts built in as there won't be as much human oversight.