Lecture 4 of 4.pdf - Google Drive

4 downloads 339 Views 93KB Size Report
Page 2 of 13. Lecture 4. • REFERENCING EXTERNAL FILES. • ODS. • LAG & RETAIN. • ARRAYS. • SAS GRAPH. •MA
Data  Processing  with  PC-­‐SAS   PubH  6325  

J.  Michael  Oakes,  PhD  

 

Associate  Professor   Division  of  Epidemiology   University  of  Minnesota   [email protected]  

  Lecture  4/4  

Lecture  4    

•   REFERENCING  EXTERNAL  FILES   •   ODS   •   LAG  &  RETAIN   •   ARRAYS   •   SAS  GRAPH   • MACROS   • STATA  

Referencing  External  Files   Indirectly  -­‐  If  you  want  to  reference  a  file  in  only  one   place  in  a  program  so  that  you  can  easily  change  it  for   another  job  or  a  later  run  use  a  FILENAME  statement     filename  mydata  ‘c:\temp\data.txt‘   filename  myprog  ‘c:\temp\prog.txt‘     data  weight;        infile  mydata;        input  idno  $  week1  week16;        loss=week1-­‐week16;     %include  myprog;  

Referencing  External  Files   Directly  -­‐   data  weight;        infile  ‘c:\temp\data.txt‘  ;        input  idno  $  week1  week16;        loss=week1-­‐week16;     %include  ‘c:\temp\prog.txt‘  ;    

ODS   The  Output  Delivery  System  (ODS)  is  a  new  “module”   in  SAS  v8.       It  permits  fancy  output  to  be  “printed”  to  the  output   window,  or  results  (eg,  stagsgcs)  to  be  output  into  a   SAS  dataset.     ODS  yields  three    types  of  fancy  output:        HTML      RTF      SAS  data  

RETAIN   Doing  calculagons  within  an  observagon  (ie,  subject)  is   relagvely  easy.  Calculagons  between  observagons  is  more   challenging!     How  many  days  since  last  visit  for  each  pagent  in  clinical  data   set.        Retain      lag      first.varname    last.varname      

Arrays   Arrays  permit  you  to  do  stuff  to  a  bunch  of  variables  at  the   same  gme.     Once  you  define  an  array  you  can  refer  to  the  array  itself,   which  stands  for  the  variable  list  inside.      

SAS  Graphs   Helpful,  but  also  consider  other  programs.       See  Dr.  Kleinman’s  tutorial  program    

Macros   Fundamentally,  macros  are  “tools”  for    text  subsgtugon.    

Macros  help:     Accomplish  repeggve  tasks  quickly  and   efficiently.     Provide  modular  structure  to  your    program.    

Macros   Macro  language  statements  begin  with  a  %     Macro  variables  start  with  a  &  

Inigalize  macro,  then  run  as  you  wish.    

Stata     Like  SAS,  but  arguably  bener:    

•   Certainly  easier  to  learn   •   Far  Less  expensive   •   Much  bener  Help  files   •   Acgve  Listserve  community   •   Super  graphics    

My  “Stata”  Course   PubH6617  

Pracgcal  Methods  for  Secondary  Data  Analysis     Fall  Term:  TuTh  8:30  am  –  9:15  am  

Lab  #4    

FINAL  EXAM