Question

我有一个看起来像这样的data.frame：

Geotype <- c(1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3)
Strategy <- c("Demand", "Strategy 1", "Strategy 2", "Strategy 3", "Strategy 4", "Strategy 5", "Strategy 6")
Year.1  <- c(1:21)
Year.2  <- c(1:21)
Year.3  <- c(1:21)
Year.4  <- c(1:21)
mydata <- data.frame(Geotype,Strategy,Year.1, Year.2, Year.3, Year.4)

我想对每年的每项策略进行总结。

这意味着我需要在数据框中的每一列下面加6行，然后跳过Demand行。然后我想对所有专栏（40年）重复这一点。

我希望输出数据框看起来像这样：

Geotype.output <- c(1, 2, 3)
Year.1.output  <- c(27, 69, 111)
Year.2.output  <- c(27, 69, 111)
Year.3.output  <- c(27, 69, 111)
Year.4.output  <- c(27, 69, 111)
output <- data.frame(Geotype.output,Year.1.output, Year.2.output, Year.3.output, Year.4.output)

有关如何优雅地做到这一点的任何建议？我尝试使用this，this和this一起破解解决方案，但我没有成功，因为我需要跳过一行。

Answer 1

您可以尝试使用base R aggregate函数（按Geotype汇总数据，使用函数sum作为“唯一值”），但使用简化的data.frame（没有“需求”行和Strategy列）：

aggregate(.~Geotype, data=mydata[mydata$Strategy !="Demand", -2], FUN=sum)
#  Geotype Year.1 Year.2 Year.3 Year.4
#1       1     27     27     27     27
#2       2     69     69     69     69
#3       3    111    111    111    111

Answer 2

使用data.table：

library(data.table)
setDT(mydata)
output = mydata[Strategy != "Demand", 
             .(Year.1.output = sum (Year.1), 
               Year.2.output = sum (Year.2), 
               Year.3.output = sum (Year.3), 
               Year.4.output = sum (Year.4)),
             by = Geotype]

#    Geotype Year.1.output Year.2.output Year.3.output Year.4.output
# 1:       1            27            27            27            27
# 2:       2            69            69            69            69
# 3:       3           111           111           111           111

我们可以通过

简化这一过程，以便更轻松地处理多年的专栏

setDT(mydata)[Strategy != "Demand", 
             lapply(.SD, sum), 
             by=Geotype, 
             .SDcols=grep("Year", names(mydata))]

Answer 3

我更喜欢以长格式获取数据：

library(dplyr)
library(tidyr)
library(reshape2)

mydata %>% gather(key, value, - Geotype, - Strategy) %>%
  filter(Strategy!="Demand") %>% group_by(Geotype, key) %>%
  summarize(sum = sum(value))

结果：

Geotype    key   sum
<dbl>  <chr> <int>
1        1 Year.1    27
2        1 Year.2    27
3        1 Year.3    27
4        1 Year.4    27
5        2 Year.1    69
6        2 Year.2    69
7        2 Year.3    69
8        2 Year.4    69
9        3 Year.1   111
10       3 Year.2   111
11       3 Year.3   111
12       3 Year.4   111

使用传播：

mydata %>% gather(key, value, - Geotype, - Strategy) %>%
  filter(Strategy!="Demand") %>% group_by(Geotype, key) %>%
  summarize(sum = sum(value)) %>% spread(key, sum)

产量

Geotype Year.1 Year.2 Year.3 Year.4
*   <dbl>  <int>  <int>  <int>  <int>
1       1     27     27     27     27
2       2     69     69     69     69
3       3    111    111    111    111

Answer 4

我的声誉太低而无法发表评论，但您可以使用dplyr和summarize_each。

mydata %>% dplyr::filter(Strategy!="Demand") %>% group_by(Geotype) %>% summarize_each(funs(sum), contains("Year"))

R对n列中的每n行求和

4 个答案: