r - 如何在 R 中合并同一数据框中的行（基于特定列下的重复值）？

Question

df 中 2 个（虚构）示例行的示例：

userid   facultyid  courseid schoolid
167       265        NA       1678  
167       71111      301      NA

假设我有几百个重复的用户 ID，就像上面的例子一样。但是，绝大多数 userid 具有不同的值。

除非第一个值为 NA （在这种情况下，NA 将重新填充来自第二个的任何值）排）？

本质上，从上面的示例中得出，我的理想输出将包含：

userid   facultyid  courseid schoolid
167       265        301       1678  

Answer 1

aggregate(x = df1, by = list(df1$userid), FUN = function(x) na.omit(x)[1])[,-1]

或使用dplyr库：

library(dplyr)

df1 %>%
  group_by(userid) %>%
  summarise_each(funs(first(na.omit(.))))

Answer 2

这是 plyr 的一个简单的单线。我写的比你问的更笼统：

 a <- data.frame(x=c(1,2,3,1,2,3,1,2,3),y=c(2,3,1,1,2,3,2,3,1),
       z=c(NA,1,NA,2,NA,3,4,NA,5),zz=c(1,NA,2,NA,3,NA,4,NA,5))

 ddply(a,~x+y,summarize,z=first(z[!is.na(z)]),zz=first(zz[!is.na(zz)]))

具体回答原始问题，如果您的数据框名为 a, ：

 ddply(a,~userid,summarize,facultyid=first(facultyid[!is.na(facultyid)]),
         courseid=first(courseid[!is.na(courseid)],
         schoolid=first(schoolid[!is.na(schoolid)])

Answer 3

这是使用的另一种方法ddply：

# requires the plyr package
library(plyr)

# Your example dataframe with added lines
schoolex <- data.frame(userid = c(167, 167, 200, 203, 203), facultyid = c(265, 71111, 200, 300, NA), 
                        courseid = c(NA, 301, 302, 303, 303), schoolid = c(1678, NA, 1678, NA, 1678))

schoolex_duprm <- ddply(schoolex, .(userid), summarize, facultyid2 = facultyid[!is.na(facultyid)][1], 
                               courseid2 = courseid[!is.na(courseid)][1], 
                               schoolid2 = schoolid[!is.na(schoolid)][1])

Answer 4

# initialize a vector that will contain row numbers which should be erased
rows.to.erase <- c()

# loop over the rows, starting from top
for(i in 1:(nrow(dat)-1)) {
  if(dat$userid[i] == dat$userid[i+1]) {
    # loop over columns to recuperate data when a NA is present
    for(j in 2:4) {
      if(is.na(dat[i,j]))
        dat[i,j] <- dat[i+1,j]
    }
    rows.to.erase <- append(rows.to.erase, i+1)
  }
}

dat.clean <- dat[-rows.to.erase,]
dat.clean
#   userid facultyid courseid schoolid
# 1    167       265      301     1678