About Reeza

Reeza · ‎11-29-2023

Don't use proc import, instead write a data step to read the files that will read each file correctly with the same types/formats. Proc import guesses, and as you see, it doesn't always make the same guess 🙂

Reeza · ‎11-29-2023

I think you want a GLM, not PROC ANOVA here. ANOVA expects the variables on the right hand side of the equal sign to be categorical.

Reeza · ‎11-29-2023

I would probably use a different variable name than execute as well, that seems like it could cause issues.

Reeza · ‎11-29-2023

First try your code without macro variables. And when using system commands, first ensure the system commands also run. So what is the issue with the code posted? Does it not work? are there errors in the log?

Reeza · ‎11-28-2023

Modifying @Tom 's code very slightly to read from the URL directly. filename src url 'https://data.cityofnewyork.us/resource/76xm-jjuj.csv'; proc format ; invalue $yntf 'Y','true'='Y' 'N','false'='N'; run; data nycems; infile src dlm=',' dsd truncover firstobs=2 ; length cad_incident_id 8 incident_datetime 8 initial_call_type $6 initial_severity_level_code 8 final_call_type $6 final_severity_level_code 8 first_assignment_datetime 8 valid_dispatch_rspns_time_indc $1 dispatch_response_seconds_qy 8 first_activation_datetime 8 first_on_scene_datetime 8 valid_incident_rspns_time_indc $1 incident_response_seconds_qy 8 incident_travel_tm_seconds_qy 8 first_to_hosp_datetime 8 first_hosp_arrival_datetime 8 incident_close_datetime 8 held_indicator $1 incident_disposition_code 8 borough $24 incident_dispatch_area $2 zipcode 8 policeprecinct 8 citycouncildistrict 8 communitydistrict 8 communityschooldistrict 8 congressionaldistrict 8 reopen_indicator $1 special_event_indicator $1 standby_indicator $1 transfer_indicator $1 ; informat incident_datetime first_assignment_datetime first_activation_datetime first_on_scene_datetime first_to_hosp_datetime first_hosp_arrival_datetime incident_close_datetime anydtdtm. zipcode comma. valid_dispatch_rspns_time_indc valid_incident_rspns_time_indc held_indicator reopen_indicator special_event_indicator standby_indicator transfer_indicator $yntf. ; format incident_datetime first_assignment_datetime first_activation_datetime first_on_scene_datetime first_to_hosp_datetime first_hosp_arrival_datetime incident_close_datetime datetime19. ; input cad_incident_id -- transfer_indicator ; run;

Reeza · ‎11-28-2023

%macro run_snno(sn_no=); data _null_; set jobs; where sn_no = &sn_no.; rc = dosubl(catt("%include('", code_name, "');")); run; %mend; One possibility. If you want to pass multiple SN_NO this could be expanded to use IN instead easily. @Nitish1003 wrote: Hi All, I am trying to create a run book to have all my code in one place with %include statement. I have a job code and code name. What I did is I created a dataset with Job_Code, code_name and added Sn_no. Now I am trying to create a macro so that I can pass the Sn_no and that particular code should get executed. Any suggestion on how should I proceed with this? Thanks a lot for looking into it

Reeza · ‎11-27-2023

Lazy way: data full_data; set ds1 ds2 indsname=src; source=src; run; proc sort data=full_data; by id window source; run; data want; set full_data; by id window source; if first.window; run; Stack the data sets identifying the source datasets Sort by ID, window and source datasets Keep ones where in DS1 first @ANKH1 wrote: I have this dataset with. Some IDs are missing values for window=0 data ds1; input ID$ window m1 m2 m3; datalines; 1 3 157 56 111 1 6 145 35 145 1 9 178 78 200 2 0 132 27 201 2 6 124 95 211 2 12 127 37 532 2 15 178 35 465 3 6 148 74 231 3 12 138 49 532 ; run; The dataset that should be used to complement data when window=0 is missing is this: data ds2; input ID$ window m1 m2 m3; datalines; 1 0 145 54 201 2 0 143 45 143 3 0 132 47 165 4 0 131 48 253 9 0 124 52 511 ; run; We need to replace missing window=0 from ds1 with values from ds2. Note that data should be added only when window=0 is missing from ds1. See the desired output below for ID 2, it keeps the data from window=0 from ds1. How can we get the below dataset? ID window m1 m2 m3 1 0 145 54 201 1 3 157 56 111 1 6 145 35 145 1 9 178 78 200 2 0 132 27 201 2 6 124 95 211 2 12 127 37 532 2 15 178 35 465 3 0 132 47 165 3 6 148 74 231 3 12 138 49 532 4 0 131 48 253 9 0 124 52 511

Reeza · ‎11-27-2023

Maybe this would be helpful https://gist.github.com/statgeek/26cf789d6da0ce941e447022fdccf0f4

Reeza · ‎11-22-2023

SAS does offer Jupyter Notebooks option to code in as well, in case that's something of interest. https://blogs.sas.com/content/sasdummy/2016/04/24/how-to-run-sas-programs-in-jupyter-notebook/ And this feature directly in SAS Studio: https://communities.sas.com/t5/SAS-Studio/Getting-the-SAS-program-Code-results-and-output-in-one-file/td-p/855789

Reeza · ‎11-22-2023

*unzip gzip file using SAS commands, requires 9.4M5+; filename target "/abc/carib/process/data/file1.sas7bdat" ; filename fromzip ZIP "/abc/123/gdr/file1.sas7bdat.gz" GZIP; data _null_; infile fromzip; file target ; input; put _infile_ ; run; libname get "/abc/carib/process/data/"; data test; set get.file1; run;

Reeza · ‎11-21-2023

Here's a solution, I expanded your sample data to include two different scenarios - one where there is a tie and one where there are missing M_Date and non missing M_date for the same ID. data have; infile cards truncover; input id 1. L_date : mmddyy10. M_date : mmddyy10.; format l_date m_date yymmdd10.;; cards; 1 9/1/2020 9/29/2020 1 8/30/2020 9/27/2020 2 8/13/2022 10/21/2022 2 10/1/2022 10/21/2022 2 10/2/2023 10/21/2022 3 12/11/2021 . 3 5/19/2022 . 4 2/5/2022 12/11/2021 4 3/19/2022 12/11/2021 4 6/21/2022 12/11/2021 5 2/5/2022 12/11/2021 5 2/5/2022 12/14/2021 5 3/19/2022 . 5 6/21/2022 . ;;; run; data step1; set have; abs_diff = abs(m_date-l_date); run; proc sort data=step1; by id abs_diff; run; data want; set step1; by id; retain flag; if first.id then flag=0; if first.id and not missing(abs_diff) then flag=1; else if flag=0 and not missing(abs_diff) then flag=1; else if flag=1 then call missing(m_date); run;

Reeza · ‎11-21-2023

And how to deal with ties?

Reeza · ‎11-21-2023

Should L be sorted in any manner? It doesn't appear to be consistently sorted.

Reeza · ‎11-20-2023

data have2; retain &varlist; set have; Array Apples_ {*} apples1-apples20; Array Carrot_{*} Carrots1-carrots20; Array oranges_{*} oranges1-oranges20; Earlyfruit=min(of apples_[*]); index_earlyfruit = whichn(earlyfruit, of apples_(*)); afruit = carrot_(index_earlyfruit); bfruit = oranges_(index_earlyfruit); run; Use WHICHN to find the index of the minimum value. Note that if you have ties, the first one is returned. @JMagenta wrote: Awesome! Thank you so much! But for the last part, where I create new variables, it gets tricky. Could this work? data have2; retain &varlist; set have; Array Apples_ {*} apples1-apples20; Array Carrot_{*} Carrots1-carrots20; Array oranges_{*} oranges1-oranges20; do i=1 to dim(apples_); Earlyfruit=min(of apples_[*]); Afruit=Carrot_[min(of apples_[*])]; /* I want the carrot that is chosen to match the suffix of Apple without changing the value of Carrot*/ Bfruit=Oranges_[min(of apples_[*])]; /* same here with Oranges*/ end; run;

Reeza · ‎11-17-2023

Assuming last 7 days. data want; set dates; in_last_7days = startdt >= intnx('month', startdt, 0, 'e') - 7; run;