Question

我有一个数据框，我想要分组，以便它有最新的数据。我必须这样做，以便我可以一次性将df插入数据库。我的数据框架如下所示：

dput(pp)
structure(list(DATE_TIME = structure(c(1369320180, 1369321980, 
1369323780, 1369325580, 1369327380, 1369285980, 1369287780, 1369289580, 
1369291380, 1369293180, 1369294980, 1369296780, 1369298580, 1369300380, 
1369302180, 1369303980, 1369305780, 1369307580, 1369309380, 1369311180
), class = c("POSIXct", "POSIXt"), tzone = ""), SITE = c("Home Page Login", 
"Home Page Login", "Home Page Login", "Home Page Login", "Home Page Login", 
"Home Page Login", "Home Page Login", "Home Page Login", "Home Page Login", 
"Home Page Login", "Home Page Login", "Home Page Login", "Home Page Login", 
"Home Page Login", "Home Page Login", "Home Page Login", "Home Page Login", 
"Home Page Login", "Home Page Login", "Home Page Login"), RESPONSE_TIME = c("0.728", 
"0.513", "0.308", "0.432", "0.877", "0.541", "0.736", "0.333", 
"0.354", "0.279", "0.178", "0.699", "0.583", "-", "0.372", "0.972", 
"1.066", "0.550", "0.246", "0.299"), AVAIL_PERCENT = c("100.00", 
"100.00", "100.00", "100.00", "100.00", "100.00", "100.00", "100.00", 
"100.00", "100.00", "100.00", "100.00", "100.00", "-", "100.00", 
"100.00", "100.00", "100.00", "100.00", "100.00"), AGENT = c(42627, 
42627, 42627, 42627, 42627, 42627, 42627, 42627, 42627, 42627, 
42627, 42627, 42627, 42627, 42627, 42627, 42627, 42627, 42627, 
42627)), .Names = c("DATE_TIME", "SITE", "RESPONSE_TIME", "AVAIL_PERCENT", 
"AGENT"), row.names = c(NA, 20L), class = "data.frame")

我构建了一个以数据框的名称作为参数的函数，一次通过数据框一个代理和一个站点，数据库检索最新的时间戳，并对数据帧进行子集化。最终的数据框将插入到数据库中。我的功能如下：

normal<-function(x) {
    x<-unique(x)
    pagelist<-unique(x$SITE)
    agentlist<-unique(x$AGENT)
    latest<-data.frame(DATE_TIME=as.POSIXct(character()), SITE=character(), RESPONSE_TIME=as.numeric(character()), AVAIL_PERCENT=as.numeric(character()), AGENT=as.numeric(character()))
for(j in 1:length(agentlist)){  
    new<-subset(x, AGENT==agentlist[j])
    for(i in 1:length(pagelist)) {

        sql1<-c("SELECT max(T.DATE_TIME) FROM <TABLE NAME> T where T.SITE=")
        sql2<-pagelist[i]
        sql<-paste(sql1, sql2, sep="'")
        sql<-gsub("\\Z","'", sql, perl=T)

        sql<-paste(sql, "and T.AGENT=", sep=" ")
        sql<-paste(sql, agentlist[j], sep="'")
        sql<-gsub("\\Z","'", sql, perl=T)
        print(sql)
        ch=odbcConnect("<userid>",pwd = "<passwd>")
        latest_date<-sqlQuery(ch, sql)
        odbcClose(ch)
        latest_date<-latest_date[1,1]
        new1<-subset(new, SITE==pagelist[i] & DATE_TIME > latest_date)
        latest<-rbind(new1, latest)
        print(head(latest,5))
        return(latest)
    }
}
}

当我这样做时：

DF＆LT; - 正常（PP）

我只获得一个网站和一个代理商。最终的df应该包含每个站点的所有代理。我在这里做错了什么想法？

Answer 1

我必须将return语句放在循环之外。

normal<-function(x) {
        x<-unique(x)
        pagelist<-unique(x$SITE)
        agentlist<-unique(x$AGENT)
        latest<-data.frame(DATE_TIME=as.POSIXct(character()), SITE=character(), RESPONSE_TIME=as.numeric(character()), AVAIL_PERCENT=as.numeric(character()), AGENT=as.numeric(character()))
    for(j in 1:length(agentlist)){  
        new<-subset(x, AGENT==agentlist[j])
        for(i in 1:length(pagelist)) {

            sql1<-c("SELECT max(T.DATE_TIME) FROM <TABLE NAME> T where T.SITE=")
            sql2<-pagelist[i]
            sql<-paste(sql1, sql2, sep="'")
            sql<-gsub("\\Z","'", sql, perl=T)

            sql<-paste(sql, "and T.AGENT=", sep=" ")
            sql<-paste(sql, agentlist[j], sep="'")
            sql<-gsub("\\Z","'", sql, perl=T)
            print(sql)
            ch=odbcConnect("<userid>",pwd = "<passwd>")
            latest_date<-sqlQuery(ch, sql)
            odbcClose(ch)
            latest_date<-latest_date[1,1]
            new1<-subset(new, SITE==pagelist[i] & DATE_TIME > latest_date)
            latest<-rbind(new1, latest)
            print(head(latest,5))
                }
    }
    return(latest)

    }

无法基于循环中的两个不同值来对数据帧进行子集化

1 个答案: