About jsmall

jsmall · ‎06-25-2019

Hi, thanks again! Works perfectly!

jsmall · ‎06-25-2019

Can I ask one quick question? I tried to make an array to do this across 5 different variables. My code is below. The only output I'm getting is for the last variable (ax3) . Any suggestions? Thank you!! data want; length leftpart $8; length rightpart $3; length ax1 ax12 axsa ax2 ax3 $8; set n6; array sub {5} axis_i_mh axis_i_mh2 axis_i_sa axis_ii axis_iii ; array x {5} x1 x2 x3 x4 x5 ; do i=1 to 5; x {5}=index(sub {5},'.'); if x {5} gt 0 then do; leftpart=substr(sub {5},1,x {5}-1); if x {5} lt length(sub {5}) then rightpart=catt('.',substr(sub{5},x{5}+1),'00'); else rightpart='.00'; end; else do; leftpart=sub{5}; rightpart='.00'; end; array sub2 {5} ax1 ax12 axsa ax2 ax3 ; do i=1 to 5; sub2 {5}=catt(leftpart,rightpart); end; end; drop leftpart rightpart x1 x2 x3 x4 x5 ; run;

jsmall · ‎06-25-2019

I don't know what's going on in those steps, but it worked! Thank you!! 🙂

jsmall · ‎06-25-2019

Hi, Thanks for your answer- I'm importing a large SAS dataset and not using the infile statement just setting it.. I tried using infile and SAS gave me a hard time. And yes, I'm trying to code this in a data step. The problem with your solution is that I have V codes, like V71.09, in this variable as well (should have mentioned that earlier) so I can't change these to numeric. So, I'm trying to figure out a way to add either no zeros, 1 or 2 zeros to character variables. Sorry, I don't have any code written yet, because I'm not sure where to start.. Thanks for any help!

jsmall · ‎06-25-2019

Hi everyone, I'm working with ICD-9 codes that were entered as text, which means clinicians could have coded certain disorders 2 ways ( i.e. 307.2 or 307.20) or 3 ways (i.e. 303, 303.0, 303.00). FYI- These are formatted as character, not numeric. In order to tabulate total number of specific diagnoses, I need these to all be in the same format. How do I add trailing zeros since only 1 is missing in some fields or 2 could be missing (meaning, I can't just multiply by 10 or 100) I've looked through the community questions here but can't find any that work in ths situation. Thanks for any help! Jen

jsmall · ‎05-15-2018

Yes! I'm having issues with the user profile! I didn't realize I was getting these error messages- any idea what's going on?: NOTE: SASUSER.PROFILE.CATALOG does not exist and is being created from NOTE: SASUSER.PROFBAK.CATALOG. NOTE: SASUSER.PROFBAK.CATALOG has been copied to SASUSER.PROFILE.CATALOG.

jsmall · ‎05-15-2018

Thanks everyone, something was going on with Windows 10, it was storing my temporary programs in my C drive, which used up all my memory. I didn't encounter these problems at all with Windows 7, so instead of working with IT who know nothing about SAS, I just decided to go back to Windows 7-seemed like the quickest, easiest solution for now.

jsmall · ‎05-10-2018

Hi there, My computer was recently upgraded to windows 10, so I changed over to SAS 9.4. Originally I had Windows 7 and was using SAS 9.3. I'm trying to import and sort a 2GB stata dataset that used to load in about 1-2 minutes with Windows7/SAS 9.3, Now it takes over 8 minutes. And then after it's imported, trying to merge this with any other dataset takes significantly more time. I'm using data steps, not proc SQL since I'm old school and haven't converted over yet 🙂 Could this be the problem? Any suggestions? Thank you! Jen

jsmall · ‎04-25-2018

Hi, I'm trying this now, I've never used this code before, so I hope it works, but it seems like SAS is going to time out, it's been churning this whole time. I'll keep you posted. Thanks for your help!

jsmall · ‎04-25-2018

Yeah, I tried exporting the variables i needed to a csv using stata and trying to import it into sas that way. there's 3 variables in the file, 2 unique ids and the character variable which is ultimately a bunch of sentences. I plan to search this text variable for key words for race, marital status, etc.. Does that help at all?

jsmall · ‎04-25-2018

Thanks for your help, but I can't get this to work. When I proc print, the narrative field comes out empty. I tried with or without @@. My file has 3 variables cid 8. Encid 8. narrative_assess_imp . I tried inputting all the variables, and that didn't work either. I think I'm confused what the a variable is in your code? Does the following log help at all? 8473 data want; 8474 infile "C:\\narrative.csv"; 8475 length narrative_assess_imp $2000; 8476 input a $char1. ; /*might need the at's might not */ 8477 if lengthn(cat(narrative_assess_imp,a)) > 2000 then do; 8478 output; 8479 narrative_assess_imp=a; 8480 end; 8481 else narrative_assess_imp=cat(narrative_assess_imp,a); 8482 run; NOTE: The infile "C:\narrative.csv" is: Filename=C:\narrative.csv, RECFM=V,LRECL=32767, File Size (bytes)=181558671, Last Modified=25Apr2018:10:06:11, Create Time=25Apr2018:10:06:08 NOTE: 397651 records were read from the infile "C:\narrative.csv". The minimum record length was 12. The maximum record length was 2061. NOTE: The data set WORK.WANT has 397651 observations and 2 variables. NOTE: DATA statement used (Total process time): real time 19.19 seconds cpu time 2.01 seconds

jsmall · ‎04-25-2018

Hi experts, I received a xlsx dataset with some text fields that are pretty massive, for example.. brief narratives or diagnosis history from a doctors visit. I can import it to stata 14 without problems, but I can't import the xlsx to SAS with either the proc import, infile or import wizard. Any idea why? And because SAS does not import Stata 14 files, even if I replace it with the older Stata version 12, it only saves those text fields to lengths of 244. Thanks for any advice or SAS code!

jsmall · ‎04-03-2018

thanks this works great!!

jsmall · ‎04-03-2018

data have; set want; input obs diab1-diab8 date2-date8; datalines; 1 . . 0 1 . . . . . . 2200 2923 2 . 0 . . 0 0 1 0 . 1496 3673 4429 5112 5962 3 . . 0 0 . . 0 1 . . 2195 2921 . . 5110 5803 4 . . 0 . . . . 0 . . 2200 . . . . 5784 5 . 0 0 1 1 1 1 1 . 1479 2200 5000 . . . . 6 . 0 0 0 0 0 . 0 . . 2168 2955 3702 . 5090 5769 7 . 0 0 0 1 1 1 1 . 1599 2199 2927 3634 4369 5118 5824 8 . . . 0 0 . 1 0 . . . 2912 3710 . 5117 5901 run; so I'm using a huge dataset, it's all in wide format. These are people who never had an official visit or diagnosis at baseline (diab1/date1) I'm was hoping to find a quick way to find out the first time (datex ) a person has a diabx of 0/1 so I can call that their baseline diagnosis. Then, if a person has a baseline diagnosis of 0 (no diabetes), and they develop diabetes afterwards, I want to find out at what time (datex ) that happened, so I can calculate their follow-up time (time to diagnosis) hope that helps. The whichn function may work?

jsmall · ‎04-03-2018

Hi, Below is longitudinal diabetes (y/n) data, I want to search for the first time a 0 or 1 appears so I can call that a baseline value. And also search for when a 1 appears so I know if the person developed diabetes. There are corresponding days2-days19 variables that are the f/u time (days from visit1) for each diabx field. So for example, obs=1242, I want to say baseline diabetes=0 occurred at the 3rd visit, and that a diabetes diagnosis was recorded at visit7, so total f/u time from baseline to diabetes diagnosis would be days7-days3. For now, we aren't worrying about the missing data between visits. I hope I explained this correctly. Thanks for any help!! Also, I don't use proc SQL very much so if it can be in a data step that would be awesome, otherwise if that's too complicated, SQL works Obs diab1 diab2 diab3 diab4 diab5 diab6 diab7 diab8 diab9 diab10 diab11 diab12 diab13 diab14 diab15 diab16 diab17 diab18 diab19 528 . . . 0 . . . . . . . . . . . . . . . 670 . . . . . . . 0 0 0 . 0 0 . 0 . 0 0 . 802 . . . . 1 . . . . . . . . . . . . . . 920 . . . . . . 0 . . . . . . . . . . . . 1148 . . . 0 0 0 . . . . . . . . . . . . . 1242 . . 0 0 0 . 1 1 . . . . . . . . . . . 1327 . . 0 0 0 0 0 0 0 0 0 0 0 . 0 . 1 . . 1898 . . . . . . . . . . . . . . . . . . . 1981 . . . . . . . . . . . . . . . . . . . 2122 . . . 0 0 . . . . . . . . . . . . . . 2124 . . . . . . . . . . . . . . . . . . . 2127 . . . . . . . . . . . . . . . . . . . 2146 . . . . . . . 0 0 0 0 0 . 1 0 1 1 . . 2936 . 0 0 0 0 0 1 1 1 1 1 1 1 . . . . . . 3059 . . . . . . 1 . . . . . . . . . . . . 3101 . 0 0 0 0 0 0 1 1 1 1 1 1 . . . . . . 3192 . . . . . . 0 0 0 0 . 0 0 0 . . 0 . . 3485 . . . . . . . 0 0 . . . . . . . . . . 3760 . 0 . 0 1 1 1 1 1 . . . . . . . . . .

Online Status	Offline
Date Last Visited	‎06-25-2019 06:47 PM

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: SAS 9.4 incredibly slow to import/sort Stata data file

Re: SAS 9.4 incredibly slow to import/sort Stata data file

SAS 9.4 incredibly slow to import/sort Stata data file

Re: import stata or xlsx file with massive text fields

Re: import stata or xlsx file with massive text fields

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: drop or add 1 or 2 trailing zeros in ICD-9 codes

drop or add 1 or 2 trailing zeros in ICD-9 codes

Re: SAS 9.4 incredibly slow to import/sort Stata data file

Re: SAS 9.4 incredibly slow to import/sort Stata data file

SAS 9.4 incredibly slow to import/sort Stata data file

Re: import stata or xlsx file with massive text fields

Re: import stata or xlsx file with massive text fields

Re: import stata or xlsx file with massive text fields

import stata or xlsx file with massive text fields

Re: Scanning variable columns in search of specific values for f/u tim...

Re: Scanning variable columns in search of specific values for f/u tim...

Scanning variable columns in search of specific values for f/u time