Início / coding / Perguntas / 79287926
Accepted
NewUsr_stat
Asked: 2024-12-17 20:46:04 +0800 CST

Uniformizar um valor de variável por outras variáveis

  • 772

suponha que tenha o seguinte:

data have;
  input ID $ (Start End) (:date.) Place $;
  format start end date9.;
cards;
0001 13JAN2015 20JAN2015 .
0001 13JAN2015 20JAN2015 1
0001 13JAN2015 20JAN2015 .
0001 21JAN2015 31DEC2015 .
0001 01JAN2018 31DEC2018 .
0001 01JAN2019 31DEC2019 .
0002 01JAN2015 31DEC2015 1
0002 01JAN2015 31DEC2015 .
0002 01JAN2015 31DEC2015 1
;

Existe uma maneira de obter o seguinte?


data have1;
  input ID $ (Start End) (:date.) Place $;
  format start end date9.;
cards;
0001 13JAN2015 20JAN2015 1
0001 13JAN2015 20JAN2015 1
0001 13JAN2015 20JAN2015 1
0001 21JAN2015 31DEC2015 .
0001 01JAN2018 31DEC2018 .
0001 01JAN2019 31DEC2019 .
0002 01JAN2015 31DEC2015 1
0002 01JAN2015 31DEC2015 1
0002 01JAN2015 31DEC2015 1
;

Em outras palavras, preciso uniformizar o valor da variável "Place" = 1 por ID e Início-Fim.

Agradeço antecipadamente

sas

2 respostas

  1. Best Answer

    Supondo que Place seja numérico:

    data have;
  input ID $ (Start End) (:date.) Place;
  format start end date9.;
cards;
0001 13JAN2015 20JAN2015 .
0001 13JAN2015 20JAN2015 1
0001 13JAN2015 20JAN2015 .
0001 21JAN2015 31DEC2015 .
0001 01JAN2018 31DEC2018 .
0001 01JAN2019 31DEC2019 .
0002 01JAN2015 31DEC2015 1
0002 01JAN2015 31DEC2015 .
0002 01JAN2015 31DEC2015 1
;

data want(drop = p);
   do until (last.End);
      set have;
      by ID Start End;
      if Place then p = 1;
   end;

   do until (last.End);
      set have;
      by ID Start End;
      Place = p;
      output;
   end;
run;

    Resultado:

    ID    Start      End        Place
0001  13JAN2015  20JAN2015  1
0001  13JAN2015  20JAN2015  1
0001  13JAN2015  20JAN2015  1
0001  21JAN2015  31DEC2015  .
0001  01JAN2018  31DEC2018  .
0001  01JAN2019  31DEC2019  .
0002  01JAN2015  31DEC2015  1
0002  01JAN2015  31DEC2015  1
0002  01JAN2015  31DEC2015  1
    • 1

  2. Uma opção seria usar PROC TIMESERIES. Defina todos os valores ausentes dentro do grupo de id start endpara o valor máximo no grupo.

    proc timeseries data=have out=want(drop=time);
    by id start end;
    var place / setmissing=max;
run;

    ID      Start       End         Place
0001    13JAN2015   20JAN2015   1
0001    13JAN2015   20JAN2015   1
0001    13JAN2015   20JAN2015   1
0001    21JAN2015   31DEC2015   .
0001    01JAN2018   31DEC2018   .
0001    01JAN2019   31DEC2019   .
0002    01JAN2015   31DEC2015   1
0002    01JAN2015   31DEC2015   1
0002    01JAN2015   31DEC2015   1
    • 0

relate perguntas