O que deu certo foi criar uma nova coluna com formato de data.   Usei o seguinte código:   PROC SQL; CREATE TABLE X AS SELECT *, INPUT(DATA_MOVIMENTO,anydtdte21.) format date9. AS ANO_MES_MOVIMENTO_1 FROM TABLE_Y ;QUIT; ... View more

How is the posted code not doing what you want?   Note that you are first removing any observations where ID is missing with the subsetting IF statement. You are then removing any extra observations per ID so that you keep only the first.   Using a subsetting IF statement before testing the FIRST.ID flag could have, in theory, caused a problem as it could have removed the observation where FIRST.ID is true.  But since you are removing all of the observations where ID is missing it doesn't really cause any trouble.   Your data step is equivalent to these other forms: data ler3; set ler2; by id; where id ne . ; if first.id then output; run; data ler3; set ler2; by id; if first.id and id ne . then output; run; data ler3; set ler2; by id; if first.id and id ne . ; run; data ler3; set ler2; by id; if id=. or not first.id then delete; run;     ... View more