Question

我有一个由以下内容组成的数据集：

ID，CATEGORY，DATE_TIME

我想删除每个ID / CATEGORY的行，这些行在任何其他记录的5分钟内都有DATE_TIME。例如，我想采取：

AAA, CAT1, 2014-12-09 18:30:58
AAA, CAT1, 2014-12-09 18:15:58
AAA, CAT1, 2014-12-09 18:12:58
AAA, CAT1, 2014-12-09 18:11:58
AAA, CAT2, 2014-12-09 18:11:58

得到这样的东西：

AAA, CAT1, 2014-12-09 18:30:58
AAA, CAT1, 2014-12-09 18:11:58
AAA, CAT2, 2014-12-09 18:11:58

感谢任何帮助！

Answer 1

加载数据，（我在5分钟后添加了一个事件，在另一个事件后添加了一秒）;

data allEvents;
    infile datalines dsd dlm=',' ;
    informat ID $3. CATEGORY $4. DATE_TIME YMDDTTM20.;
    format DATE_TIME DATETIME19.2;
    input ID $ CATEGORY $ DATE_TIME ;
    datalines;
AAA, CAT1, 2014-12-09 18:30:58
AAA, CAT1, 2014-12-09 18:16:59
AAA, CAT1, 2014-12-09 18:15:58
AAA, CAT1, 2014-12-09 18:12:58
AAA, CAT1, 2014-12-09 18:11:58
AAA, CAT2, 2014-12-09 18:11:58
;
run;

在ID，CATEGORY和DATE_TIME ;

上对其进行排序

proc sort data=allEvents;
    by ID CATEGORY DATE_TIME;
run;

在数据步骤中读取并过滤;

data wantedEvents (drop=writtenStamp);
    set allEvents;
    by ID CATEGORY DATE_TIME;

    ** remember the last written DATE_TIME **;
    retain writtenStamp;

    if first.CATEGORY then do;
        output;
        writtenStamp = DATE_TIME;
    end;
    else if DATE_TIME GT writtenStamp + hms(0,5,0) then do;
        output;
        writtenStamp = DATE_TIME;
    end;
run;

按原始顺序对其进行排序;

proc sort data=wantedEvents;
    by ID CATEGORY descending DATE_TIME ;
run;

在X分钟内删除行

1 个答案: