About michtka

michtka · ‎04-25-2012

data_null...in termns of using proc means instead of proc summary...is not possible to get the same result...can you give me some suggestion of the formatting using proc means?...Thanks.

michtka · ‎04-25-2012

sorry, I need to try all of this first...but my last question is, if I use proc means instead of proc univariate, proc means is following the same syntax, i.e i can get the same result using proc means? Thanks.

michtka · ‎04-23-2012

Hi, using this ods outputs, you get all the moments of all the phases: phase1, phase2 and phase3. ods listing close; ods output moments=ustat (keep=varname treat visit label1 nvalue1 where=(label1 in : ('N' 'Mean' 'Std'))); proc univariate data=new; var phase: ; class treat visit; run; ods output close; ods listing; but if I use this other statement I only get the moments of the phase 1: proc univariate data=new; var phase: ; class treat visit; output out=ustat2 n=n mean=mean std=std; run; Why?

michtka · ‎04-20-2012

Hi Everyone, can someone help me to give some information (examples) how apply the Bayesian Model Selection and PCA analysis in the pharma industry? I have knowledge about these two multivariate analysis and i would like to know more about their application in the pharmaceutical industry. Thanks.

michtka · ‎04-19-2012

data new; input visit &$6. treat &$7. pop :1. phase1 :1. phase2 :1. phase3 :1. dup :1. num :1.; cards; week12 placebo 1 1 1 2 1 1 week12 placebo 1 1 1 2 1 2 week12 placebo 1 1 1 2 1 3 week28 placebo 1 2 2 1 2 4 week28 placebo 1 2 2 1 2 5 week28 placebo 1 2 2 1 2 6 week40 placebo 1 3 3 2 3 7 week40 placebo 1 3 3 2 3 8 week40 placebo 1 3 3 2 3 9 week64 placebo 1 4 3 1 4 10 week64 placebo 1 4 3 1 4 11 week64 placebo 1 4 3 1 4 12 week12 drug 2 4 3 1 5 13 week12 drug 2 4 3 1 5 14 week12 drug 2 4 3 1 5 15 week40 drug 2 3 5 1 6 16 week40 drug 2 3 5 1 6 17 week40 drug 2 3 5 1 6 18 week64 drug 2 4 3 4 7 19 week64 drug 2 4 3 4 7 20 week64 drug 2 4 3 4 7 21 week12 placebo 3 3 3 1 8 22 week12 placebo 3 3 3 1 8 23 week12 placebo 3 3 3 1 8 24 week28 placebo 3 4 1 3 9 25 week28 placebo 3 4 1 3 9 26 week28 placebo 3 4 1 3 9 27 week64 placebo 3 1 3 4 10 28 week64 placebo 3 1 3 4 10 29 week64 placebo 3 1 3 4 10 30 week12 drug 4 5 4 4 11 31 week12 drug 4 5 4 4 11 32 week12 drug 4 5 4 4 11 33 week28 drug 4 3 3 5 13 34 week28 drug 4 3 3 5 13 35 week28 drug 4 3 3 5 13 36 week40 drug 4 5 3 4 14 37 week40 drug 4 5 3 4 14 38 week40 drug 4 5 3 4 14 39 week64 drug 4 1 1 2 15 40 week64 drug 4 1 1 2 15 41 week64 drug 4 1 1 2 15 42 week12 placebo 5 2 1 4 16 43 week12 placebo 5 2 1 4 16 44 week12 placebo 5 2 1 4 16 45 week28 placebo 5 2 2 1 17 46 week28 placebo 5 2 2 1 17 47 week28 placebo 5 2 2 1 17 48 ; run; proc sort data=new out=new2; by dup; run; data new2s; set new2; by dup; if first.dup then output; run; proc sort data=new2s out=new2ss; by treat; run; proc transpose data=new2ss out=new2t (rename=(_name_=phasex)); by treat dup visit; var phase1 phase2 phase3; run; data new2cat; set new2t; value=cats(visit,phasex); run; proc sort data=new2cat out=new2cats; by treat value; run; proc means data=new2cats noprint; by treat value; output out=newxxx n=n; var col1; run; proc transpose data=newxxx out=new2catst; by treat; id value; var n; run;

michtka · ‎04-19-2012

Hi guys, can anyone help me with this: I got the next dataset new: data new; input visit &$6. treat &$7. pop :1. phase1 :1. phase2 :1. phase3 :1.; cards; week12 placebo 1 1 1 2 week28 placebo 1 2 2 1 week64 placebo 1 1 4 2 week40 placebo 1 2 4 1 week40 drug 1 3 5 1 week12 drug 1 4 3 1 week12 drug 1 3 3 1 week28 placebo 1 3 2 2 week28 placebo 1 4 1 3 week40 placebo 1 4 3 4 week40 placebo 1 5 3 4 week12 drug 1 5 4 4 week64 drug 1 1 3 4 week64 drug 1 3 3 5 week12 drug 1 3 2 3 week12 placebo 1 2 1 4 week64 placebo 1 1 1 2 week28 placebo 1 2 2 1 ; run; I would like something like that: week12 week28 week40 week64 phase1 phase2 phase3 phase1 phase2 phase3 phase1 phase2 phase3 phase1 phase2 phase3 placebo N x x x x x x x x x x x x mean x x x x x x x x x x x x sd x x x x x x x x x x x x drug N x x x x x x x x x x x x mean x x x x x x x x x x x x sd x x x x x x x x x x x x I start trying do something like this: proc sort data=new out=new2; by treat visit; run; proc means data=new2 noprint; by treat visit; output out=new3 n=n mean=mean std=std; var phase1 phase2 phase3; run; but it doesnt work.....any help? Thanks in advance. V

michtka · ‎04-19-2012

Thanks Astounding for your help, I will keep your piece of code in my mind, as well as the execute function proposed by Ksharp (I dont need it at the moment). Only just, I needed to be more "efficient" (I known is not the right word) format, as Ksharp suggested.

michtka · ‎04-19-2012

Thanks Ksharp, my final code with your help become like this (in boldface are the changes with my original code). data new; input visit &$6. treat &$7. pop :1. phase &$2.; cards; week12 placebo 1 I week28 placebo 1 II week40 drug 1 III week64 drug 1 IV week64 placebo 1 III week40 placebo 1 III week40 drug 1 I week12 drug 1 II week12 drug 1 II week28 placebo 1 II week28 placebo 1 II week40 placebo 1 III week40 placebo 1 IV week12 drug 1 IV week64 drug 1 III week64 drug 1 II week12 drug 1 I week12 placebo 1 I week64 placebo 1 II week28 placebo 1 III ; run; *summary of subjects by phase; proc summary data=new nway; class treat phase visit; var pop; output out=newsum(drop=_:) n=n; run; *summary of subjects by treat; proc summary data=new nway; class treat; var pop; output out=new0sum (drop=_:) n=n2; run; *merging the summaries; data newsumtotal; merge new0sum newsum; by treat; length value $ 20; value=catx(' ',n,cats('(',put(n/n2,percent8.2),')')); run; *transposing to get the final table; proc transpose data=newsumtotal out=newsumtotalt; by treat n2 phase; id visit; var value; run;

michtka · ‎04-18-2012

Thanks Hai, I will use the first one with retain.

michtka · ‎04-18-2012

Hi everyone, i got the next dataset: visit phase baseline 1 week2 2 week4 2 week6 1 baseline 2 week2 1 week4 1 week6 1 I want something like this visit phase visit_b phase_b baseline 1 baseline 1 week2 2 baseline 1 week4 2 baseline 1 week6 1 baseline 1 baseline 2 baseline 2 week2 1 baseline 2 week4 1 baseline 2 week6 1 baseline 2 Any help...maybe with the retain function...thanks in advance. V

michtka · ‎04-18-2012

Thanks Ksharp. I couple of things: 1. You use proc freq intead of proc summary...it is another option as we talked. good. 2. Very nice the option value making the right format before transpose the varibales. Thank you. (I will get that the right answer). 3. Thank you for the last part, making the code more standard using the libraries of SAS, and using the function execute. At the moment I dont need it because (from the transpose of visit, the variables weekX are generated, and I just got your solution in point 2, making the formatting before the transpose) Thanks!

michtka · ‎04-17-2012

Hello guys, anyone can help me to make this code more efficient(see code below), please not using sql: I would like to do something with the week12, week28, week40 and week64, try to do it more standard, because if I need to change the dataset new with different weeks (i.e week1, wek3, week5, week52), I would have to change manually these weeks..and it is no really efficient. something like week_i,where i=1 to 4 Any help? Thanks in advance. Please see the code below: data new; input visit &$6. treat &$7. pop :1. phase &$2.; cards; week12 placebo 1 I week28 placebo 1 II week40 drug 1 III week64 drug 1 IV week64 placebo 1 III week40 placebo 1 III week40 drug 1 I week12 drug 1 II week12 drug 1 II week28 placebo 1 II week28 placebo 1 II week40 placebo 1 III week40 placebo 1 IV week12 drug 1 IV week64 drug 1 III week64 drug 1 II week12 drug 1 I week12 placebo 1 I week64 placebo 1 II week28 placebo 1 III ; run; *summary of subjects by phase; proc summary data=new nway; class treat phase visit; var pop; output out=newsum(drop=_:) n=n; run; *summary of subjects by treat; proc summary data=new nway; class treat; var pop; output out=new0sum (drop=_:) n=n2; run; *merging the summaries; data newsumtotal; merge new0sum newsum; by treat; run; *transposing to get the final table; proc transpose data=newsumtotal out=newsumtotalt; by treat n2 phase; id visit; var n; run; *convert numeric variable to character (I want to make this part more efficient); data newf (drop=week12 week28 week40 week64); set newsumtotalt; _week12=put(week12,3.0)||' '||'('||put(week12/n2,percent6.2)||')'; _week28=put(week28,3.0)||' '||'('||put(week28/n2,percent6.2)||')'; _week40=put(week40,3.0)||' '||'('||put(week40/n2,percent6.2)||')'; _week64=put(week64,3.0)||' '||'('||put(week64/n2,percent6.2)||')'; run; data newfr (rename=(_week12=week12 _week28=week28 _week40=week40 _week64=week64) drop=_name_ ); retain treat n2 phase _week12 _week28 _week40 _week64; set newf; run; V

michtka · ‎04-13-2012

thnaks, I did with n

michtka · ‎04-13-2012

I did it adding missing as an option of proc summary...I think will be the same.

michtka · ‎04-13-2012

I got it proc summary data=new nway missing; class treattxt eff_rslt visder; var itt; output out=newsum(drop=_:) n=n; run;

Online Status	Offline
Date Last Visited	‎09-01-2015 07:11 AM

Re: HR interpretation for 3 different scenarios

Re: HR interpretation for 3 different scenarios

Re: HR interpretation for 3 different scenarios

Re: HR interpretation for 3 different scenarios

Re: HR interpretation for 3 different scenarios

HR interpretation for 3 different scenarios

get then minimum value when you have missing values

Re: macro (Warning:apparent symbol not resolved)

Re: macro (Warning:apparent symbol not resolved)

macro (Warning:apparent symbol not resolved)

Re: show row with missing data in a summary table

Re: code more efficient

Re: removing leading zeros of a character variable

Index where if

Re: multimacrovariables in sql

Re: sorting dataset descending frequency

Re: proc sql instead of proc sort nodupkey

Re: retain the format var in proc transpose

Re: proc transpose 4 variables

Re: proc transpose 4 variables

Bayesian Model Selection and PCA

Re: proc transpose 4 variables

proc transpose 4 variables

Re: code more efficient

Re: code more efficient

Re: retain variables for baseline

retain variables for baseline

Re: code more efficient

code more efficient

Re: proc summary nway help

Re: proc summary nway help

Re: retain the format var in proc transpose