【问题标题】:Sum values from one df base on two columns of another df基于另一个 df 的两列对来自一个 df 的值求和
【发布时间】:2021-02-23 22:37:31
【问题描述】:

这些是两个数据帧的子集。

我有一个 df1,每个地块一年中的每一天都有 degree_days 的测量值:

date plot degree_days
5-13-19 1 3.5
5-13-19 2 5.35
5-13-19 3 4.8
5-14-19 1 4.5
5-14-19 2 4.4
5-14-19 3 5.8
5-15-19 1 3.5
5-15-19 2 5.35
5-15-19 3 4.8
5-16-19 1 4.5
5-16-19 2 4.4
5-16-19 3 5.8

(续)

还有一个 df2,每个图中第一朵花的平均日期:

mean_first_flower_date plot
5-16-19 1
8-5-19 2
6-12-19 3
5-16-19 1
8-5-19 2
6-12-19 3
5-16-19 1
8-5-19 2
6-12-19 3

(续)

我的目标是使用 df1 在 df2 中创建一个新变量,该变量将在每个图中将 degree_days 与 mean_first_flower_date 相加。例如,如果在 df2 中,地块 1 中的 mean_first_flower_date 为 5-16-19,则新变量为 16(即 3.5 + 4.5+ 3.5 + 4.5 的总和)"

mean_first_flower_date plot new_var
5-16-19 1 16

我尝试了这个 for 循环,但我不知道如何根据每个图来调整总和:


for (i in 1:nrow(df2)) {
    df2[i, "mean_first_flower_date"] -> flower 
    print(i)
    df1[df1$date <= flower, ] -> df1.temporary 
    sum(df1.temporary$degree_days) -> new_var 
    df2[i, "new_var"] <- new_var 
}

非常感谢任何帮助!

谢谢!

【问题讨论】:

    标签: r for-loop


    【解决方案1】:

    这是通过定义用户函数f 的基本R 选项(感谢@akrun 的数据)

    f <- function(d, p) with(subset(df1, date <= d & plot == p), sum(degree_days))
    
    dfout <- within(
      df2,
      new_var <- Vectorize(f)(mean_first_flower_date, plot)
    )
    

    这样

    > dfout
      mean_first_flower_date plot new_var
    1             2019-05-16    1    16.0
    2             2019-08-05    2    19.5
    3             2019-06-12    3    21.2
    4             2019-05-16    1    16.0
    5             2019-08-05    2    19.5
    6             2019-06-12    3    21.2
    7             2019-05-16    1    16.0
    8             2019-08-05    2    19.5
    9             2019-06-12    3    21.2
    

    【讨论】:

      【解决方案2】:

      将“日期”列转换为Date 类后,非等值连接会很有用

      library(data.table)
      new_var <- setDT(df1)[df2, sum(degree_days), on =
               .(plot, date <= mean_first_flower_date), by = .EACHI]$V1
      setDT(df2)[, new_var := new_var][]
      

      -输出

      df2
      #   mean_first_flower_date plot new_var
      #1:             2019-05-16    1    16.0
      #2:             2019-08-05    2    19.5
      #3:             2019-06-12    3    21.2
      #4:             2019-05-16    1    16.0
      #5:             2019-08-05    2    19.5
      #6:             2019-06-12    3    21.2
      #7:             2019-05-16    1    16.0
      #8:             2019-08-05    2    19.5
      #9:             2019-06-12    3    21.2
      

      数据

      df1 <- structure(list(date = c("5-13-19", "5-13-19", "5-13-19", "5-14-19", 
      "5-14-19", "5-14-19", "5-15-19", "5-15-19", "5-15-19", "5-16-19", 
      "5-16-19", "5-16-19"), plot = c(1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 
      3L, 1L, 2L, 3L), degree_days = c(3.5, 5.35, 4.8, 4.5, 4.4, 5.8, 
      3.5, 5.35, 4.8, 4.5, 4.4, 5.8)), class = "data.frame", row.names = c(NA, 
      -12L))
      
      df2 <- structure(list(mean_first_flower_date = c("5-16-19", "8-5-19", 
      "6-12-19", "5-16-19", "8-5-19", "6-12-19", "5-16-19", "8-5-19", 
      "6-12-19"), plot = c(1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L)), 
      class = "data.frame", row.names = c(NA, 
      -9L))
      
      df1$date <- as.Date(df1$date, "%m-%d-%y")
      df2$mean_first_flower_date <- as.Date(df2$mean_first_flower_date, "%m-%d-%y")
      

      【讨论】:

        猜你喜欢
        • 2018-04-04
        • 1970-01-01
        • 2016-02-29
        • 2019-11-02
        • 2020-01-25
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        相关资源
        最近更新 更多