Question

我有一个大型数据集，每个电台具有相同的纬度和经度。在数据集中，一些行缺少lat和lon，而是说'unknown'。我需要在没有丢失数据的其他站点填写lat long的未知数。

在这个例子中，我希望第5行为lat和lon插入3和8：

> station <- c("a","b","c","c","c")
> lat <- c("1","2","3","3","unknown")
> lon <- c("6","7","8","8","unknown")
> data.frame(station,lat,lon)
  station     lat     lon
1       a       1       6
2       b       2       7
3       c       3       8
4       c       3       8
5       c unknown unknown

我的数据集中有一百万行，如果需要几分钟才能完成，那很好，因为这只会在分析开始前运行一次。除非真的有必要，否则我宁愿不安装另一个包。

Answer 1

这样的事，也许 -

df$station <- as.character(df$station)

unknownstations <- unique(subset(df,df$lat == "unknown","station"))
unknownstationscoords <- unique(subset(df,station %in% unknownstations$station & lat != "unknown"))

for( i in unknownstations$station)
{
df[df$station == i,"lat"] <- subset(unknownstationscoords,station %in% i,"lat")
df[df$station == i,"lon"] <- subset(unknownstationscoords,station %in% i,"lon")
}

Answer 2

我在动物园包中使用na.locf。首先，我将unknown更改为NA，然后应用na.locf：

> library(zoo)
> df[ df=="unknown"] <- NA
> df2 <- do.call(rbind, lapply(split(df, df$station), na.locf))
> df2[, -1]  <- sapply(df2[, -1], as.numeric)  # numeric variables should be numeric 
> df2
    station lat lon
a         a   1   6
b         b   2   7
c.3       c   3   8
c.4       c   3   8
c.5       c   3   8

如果您想要使用rownames，请使用rownames并指定名称：

> rownames(df2) <- 1:nrow(df2)
> df2
  station lat lon
1       a   1   6
2       b   2   7
3       c   3   8
4       c   3   8
5       c   3   8

Answer 3

y=function(station,lat,lon){

  temp=cbind(station,lat,lon)
  lat_ind=lat!="unknown"
  lon_ind=lon!="unknown"


  if(all(lat_ind)==0){
    hash=unique(temp[lat_ind,])
    ind2=hash[,1]==station[!lat_ind]
    temp[!lat_ind,]=temp[ind2,]

    return(temp) 

  }else if(all(lon_ind)==0){
    hash=unique(temp[lon_ind,])
    ind2=hash[,1]==station[!lon_ind]
    temp[!lon_ind,]=temp[ind2,]

    return(temp)


  }else {

    return(temp)
  }


}




##case1

station <- c("a","b","c","c","c")
lat <- c("1","2","3","3","unknown")
lon <- c("6","7","8","8","unknown")

y(station,lat,lon)
# station lat lon
# [1,] "a"     "1" "6"
# [2,] "b"     "2" "7"
# [3,] "c"     "3" "8"
# [4,] "c"     "3" "8"
# [5,] "c"     "3" "8"


##case2

station <- c("a","b","c","c","c")
lat <- c("1","2","3","3","3")
lon <- c("6","7","8","8","unknown")
y(station,lat,lon)
# station lat lon
# [1,] "a"     "1" "6"
# [2,] "b"     "2" "7"
# [3,] "c"     "3" "8"
# [4,] "c"     "3" "8"
# [5,] "c"     "3" "8"


##case3

station <- c("a","b","c","c","c")
lat <- c("1","2","3","3","unknown")
lon <- c("6","7","8","8","8")
y(station,lat,lon)
# station lat lon
# [1,] "a"     "1" "6"
# [2,] "b"     "2" "7"
# [3,] "c"     "3" "8"
# [4,] "c"     "3" "8"
# [5,] "c"     "3" "8"

R数据框根据其他数据填充缺失值

3 个答案: