Question

我的表格如下：

dt<-data.frame(Date=c("2011-01-16","2011-01-16","2011-07-08","2011-07-09","2011-07-09","2011-08-17","2011-09-10","2011-09-11","2011-09-11"),Number=c(7,7,NA,1,1,NA,7,5,6),Hour=c(0.25,0.25,NA,0.6,0.6,NA,2,0.25,0.25))
        Date Number Hour
1 2011-01-16      7 0.25
2 2011-01-16      7 0.25
3 2011-07-08     NA   NA
4 2011-07-09      1 0.60
5 2011-07-09      1 0.60
6 2011-08-17     NA   NA
7 2011-09-10      7 2.00
8 2011-09-11      5 0.25
9 2011-09-11      6 0.25

我想用Hour和Number对Date求和。输出看起来像这样：

      Date    "1"  "5" "6"  "7"
1 2011-01-16  NA   NA   NA 0.5
2 2011-07-08  NA   NA   NA  NA
3 2011-07-09 1.2   NA   NA  NA
4 2011-08-17  NA   NA   NA  NA
5 2011-09-10  NA   NA   NA 2.0
6 2011-09-11  NA 0.25 0.25  NA

您能建议我一个获取输出的函数吗？

Answer 1

您可以为此使用聚合函数。

dt$Date <- as.character(dt$Date)
aggregate(dt$Hour, by = list(dt$Number, dt$Date), FUN = function(x) sum(x, na.rm = T))

或者，您可以使用它（这次不删除NA）：

with(dt, aggregate(Hour, by = list(Number, Date), FUN = sum))

Answer 2

我们可以利用fun.aggregate中的dcast

library(data.table)
dcast(setDT(dt), Date + Hour ~ Number, sum)

如果OP打算在没有组合的情况下获得NA，则创建一个条件，因为sum的长度为0（sum(integer(0))）返回0

dcast(setDT(dt), Date + Hour ~ Number, function(x) 
    if(length(x) == 0) NA_real_ else sum(x, na.rm = TRUE))[,
          .(Date, Hour, `1`, `5`, `6`, `7`)]
#.       Date Hour   1    5    6   7
#1: 2011-01-16 0.25  NA   NA   NA 0.5
#2: 2011-07-08   NA  NA   NA   NA  NA
#3: 2011-07-09 0.60 1.2   NA   NA  NA
#4: 2011-08-17   NA  NA   NA   NA  NA
#5: 2011-09-10 2.00  NA   NA   NA 2.0
#6: 2011-09-11 0.25  NA 0.25 0.25  NA

Answer 3

我们可以为每个组group_by Date和Number和sum Hour使用spread将其更改为宽格式。但是，这也会提供NA列（因为Number具有NA值），如果不需要，可以将其删除。

library(dplyr)

dt %>%
  group_by(Date, Number) %>%
  summarise(Hour = sum(Hour, na.rm = TRUE)) %>%
  tidyr::spread(Number, Hour) %>%
  select(-`<NA>`)

#  Date         `1`   `5`   `6`   `7`
#  <fct>      <dbl> <dbl> <dbl> <dbl>
#1 2011-01-16  NA   NA    NA      0.5
#2 2011-07-08  NA   NA    NA     NA  
#3 2011-07-09  1.2  NA    NA     NA  
#4 2011-08-17  NA   NA    NA     NA  
#5 2011-09-10  NA   NA    NA      2  
#6 2011-09-11  NA   0.25  0.25  NA

按日期求和R中的行

3 个答案: