Home

curious_tad

Hello everybody, fairly new in the world of SAS, a curious person,say, learning in his own time. I am trying to riddle out how to make a successful POST/GET request to REST API using %macro and JSON as the primary payload and then set up a table that tracks a history of past requests and responses. Set this task to myself as it's one of the more common things people aim to do in other languages. Working with APIs. While, it takes me about 10 minutes to write such logic in JS or Python without any headaches, though I cannot string up a working logic in SAS using %macro (a term function could also be used as a synonym I guess?). A non-macro approach clicked using datalines and a hard-coded value as a prompt. I am now well aware that datalines will not budge when using macro logic so I went with put, but still not a single budge. Here's the code that I have managed to jot down so far. The API is crudcrud.com allowing up to 100 free requests. %macro api_post_request(prompt=); %let api_url=url; filename json_in temp; data _null_; file in; put "{ ""username"":""&prompt"" }" ; run; filename resp temp; proc http method="POST" url="&api_url ct="application/json" in=json_in out=resp; headers "Content-Type"="application/json"; run; %mend; %api_post_request(prompt=Player1); Fierce thanks!

Jinlong_SUFE

Hi Guys, I am trying to obtain the median value using median() function in PROC SQL as follows (SAS9.4): /* select clause with additional columns: Name Height*/ proc sql; select Name,Sex,Height,median(Height) as Med_Height,mean(Height) as Mean_Height from sashelp.class group by sex having sex="M"; quit; The results are as follows: However, when removing columns Name, Height in select clause like this: /* select clause without additional column */ proc sql; select Sex, median(Height) as Med_Height, mean(Height) as Mean_Height from sashelp.class group by sex having sex="M"; quit; The results are as follows: Comparing the two tables, the mean value of Height is still 63.91 while the median of Height changes from 64.3 to 64.15 (of course the 64.15 is the right value). Where did the first median value 64.3 come from? How did this difference (64.3 v.s. 64.15) happen? After all the mean value remains the same. I feel confused. Can anyone explain this? Thanks a lot guys. Appreciate it.

JB1_DK

Been a long-time SAS user and wanted to try out CAS. I came across the concept of a Personal CAS Server which is like a Sandbox environment. So seems perfect for my process right now. But there is surprisingly no information to be found on how to set it up. So how do I set up a Personal CAS Server - does anyone know ?

KKIND

proc genmod data=NIS.nis_2019_septicemia_ipexcmicat; class pay1recat (ref = "3") amonth (ref = "1") aweekend (ref = "0") died (ref = "0") elective (ref = "0") female (ref = "0") zipinc_qrtl (ref = "4") hosp_bedsize (ref = "3") hosp_locteach (ref = "1") hosp_region (ref = "1") aprdrg_risk_mortality (ref = "0") aprdrg_severity (ref = "0") pl_nchsrecat (ref = "1") racerecat (ref = "4") cmr_index_mortalitycat (ref = "1") / param=ref; model los = pay1recat age pay1recat*age amonth pay1recat*amonth aweekend died pay1recat*died elective pay1recat*elective female pay1recat*female i10_ndx pay1recat*i10_ndx i10_npr pay1recat*i10_npr zipinc_qrtl pay1recat*zipinc_qrtl hosp_bedsize pay1recat*hosp_bedsize hosp_locteach pay1recat*hosp_locteach hosp_region pay1recat*hosp_region aprdrg_risk_mortality pay1recat*aprdrg_risk_mortality aprdrg_severity pay1recat*aprdrg_severity pl_nchsrecat pay1recat*pl_nchsrecat racerecat pay1recat*racerecat cmr_index_mortalitycat pay1recat*cmr_index_mortalitycat totchg pay1recat*totchg / dist=negbin link=log; run; Log WARNING: The negative of the Hessian is not positive definite. The convergence is questionable. WARNING: The procedure is continuing but the validity of the model fit is questionable. WARNING: The specified model did not converge. WARNING: Negative of Hessian not positive definite.

SAS_USER_928

Hi, I need some help with writing a SAS code to handle the following scenario. I have a data set that has multiple sites; each site has multiple patients; and each patient can have data of up to 7 days (one row represents one day). The data set has form start and save time. Data set is also setup to have no duplicates. First day data of each patient should always have the previous save and open time (both are user defined variables) as NULL. e.g., sr.1, 8, and 9 will have previous open and save time variables as NULL. And using these variables, I want to calculate the the following for each patient at a site: number of hours a patient took between today's and yesterday's start times (i.e., from today's start to previous start) and number of hours a patient took between today's open and yesterday's start times (i.e., from today's start to previous day save) Sr. PT SITE START_TIME SAVE_TIME 1 001 01 4/9/2024 16:44 4/9/2024 16:46 2 001 01 4/10/2024 16:39 4/10/2024 16:40 3 001 01 4/11/2024 16:07 4/11/2024 16:08 4 001 01 4/12/2024 16:05 4/12/2024 16:06 5 001 01 4/13/2024 20:05 4/13/2024 20:06 6 001 01 4/14/2024 19:31 4/14/2024 19:52 7 001 01 4/15/2024 16:25 4/15/2024 16:25 8 002 01 4/15/2024 16:19 4/15/2024 16:20 9 003 01 4/15/2024 16:21 4/15/2024 16:22 10 002 02 4/3/2024 18:34 4/3/2024 18:41 11 002 02 4/4/2024 18:03 4/4/2024 18:09 12 002 02 4/6/2024 21:11 4/6/2024 21:14 13 002 02 4/6/2024 21:14 4/6/2024 21:16 14 002 02 4/7/2024 18:17 4/7/2024 18:20 15 002 02 4/8/2024 18:12 4/8/2024 18:14 16 002 02 4/9/2024 18:06 4/9/2024 18:08 17 004 02 4/5/2024 18:03 4/5/2024 18:10 18 004 02 4/6/2024 16:27 4/6/2024 16:32 19 004 02 4/7/2024 19:18 4/7/2024 19:22 20 004 02 4/8/2024 18:01 4/8/2024 18:07 21 004 02 4/9/2024 18:06 4/9/2024 18:16 22 004 02 4/10/2024 18:08 4/10/2024 18:10 23 004 02 4/11/2024 19:22 4/11/2024 19:26 24 005 02 4/8/2024 18:05 4/8/2024 18:07 25 005 02 4/9/2024 19:32 4/9/2024 19:34 26 005 02 4/10/2024 18:45 4/10/2024 18:47 27 005 02 4/11/2024 18:02 4/11/2024 18:04 28 005 02 4/12/2024 18:04 4/12/2024 18:07 29 005 02 4/13/2024 18:01 4/13/2024 18:03 30 005 02 4/14/2024 18:06 4/14/2024 18:08 Thanks so much in advance for all the help.