将值赋给R中的特定列和行

我有一个数据框，如下所示

Name Amount Subscriptionperiod(Months)  Subscriptionstart (Month)
Tom  300    3                             0
Tom  100    3                             2
Jim  500    5                             0
Jim  600    3                             1

我想安排如下数据。例如，Tom在一笔交易中支付了300美元，为期3个月。在2个月后的第二笔交易中，他在3个月内额外支付了100美元

吉姆也一样

Name   M0   M1  M2   M3  M4  M5  M6
Tom    300 300  300  0   0   0   0
Tom    0    0   100  100 100 0   0
Jim    500 500  500  500 500 0   0
Jim    0   600  600  600 0   0   0

我不能改变。我使用下面的代码来完成第一部分。但是对于Jim，如何创建第二行，其中值从M2开始。100美元，从M2开始，持续到M4

for(i in 0:6) df <- within(df,assign(paste0("M",i),ifelse((Subscriptionperiod>i),amount,0)))

---

首先，让我们从最小数据帧开始：

df1 <- data.frame(name=c("Tom", "Tom", "Jim", "Jim"), amount=c(300, 100, 500, 600), 
                  Subperiod=c(3, 3, 5, 3), SubStart = c(0, 2, 0, 1))

> df1
  name amount Subperiod SubStart
1  Tom    300         3        0
2  Tom    100         3        2
3  Jim    500         5        0
4  Jim    600         3        1

现在，聪明的部分是创建一个函数，根据您的规则创建一个大向量来填充矩阵

special_spread <- function(df1){
    bigrow <- c()
    for(i in 1:nrow(df1)){
      pt1 <- rep(0, df1$SubStart[i])
      pt2 <- rep(df1$amount[i], df1$Subperiod[i])
      pt3 <- rep(0, ncol(m) - (length(pt2)+length(pt1)) )
      bigrow <- c(bigrow, pt1, pt2, pt3)
    }
    m1 <- as.data.frame(matrix(bigrow, nrow=4, ncol=7, byrow = TRUE))
    m1 <- cbind(df1$name, m1)
    colnames(m1) <- c("name", paste0("M", 0:6))
    return(m1)
}

> special_spread(df1)
  name  M0  M1  M2  M3  M4 M5 M6
1  Tom 300 300 300   0   0  0  0
2  Tom   0   0 100 100 100  0  0
3  Jim 500 500 500 500 500  0  0
4  Jim   0 600 600 600   0  0  0

---

请告诉我这是否需要更多的解释，如果这或多或少回答了您的问题。

我使用data.table和plyr包来实现这一点

library(data.table)
library(plyr)

df <- data.table(read.table(text='Name Amount Period  Start
                                   Tom  300    3      0
                                   Tom  100    3      2
                                   Jim  500    5      0
                                   Jim  600    3      1', header=T, row.names = NULL))
#Create a row by repeating df$Amount, df$Period times and padding with 0
create_rows <- function(x, y){
  c(rep(0, x$Start), rep(x$Amount, x$Period), rep(0, y - x$Period - x$Start))
}

#Create a new data.table and add Name column
df2 <- data.table(Name = df$Name)

#Create an array of month names
months <- paste('M', 0:6, sep = '')

#Use adply (from plyr) to apply create_rows() accross all rows of df2
#.expand = FALSE ensures the size of the returned data.frame is the right size
#.id = NULL stops adply() from creating an index column
#with = FALSE allows the variable month to be used to refer to columns
df2[, months := adply(df, 
                      1, create_rows,
                      length(months), 
                      .expand = FALSE, 
                      .id =NULL),
    with = FALSE]

---

您可以使用df3[is.nadf3]将所有NAs更改为0，而不是使用paste…，sep=，您也可以使用paste0…，这在数据帧中有效吗？因为我的源是一个框架，所以需要首先使用setDTdf将数据转换为data.table。Data.table是Data.frame的扩展。除少数例外情况外，数据帧操作仍能正常工作，如有必要，可以通过setDF进行简单的转换。@NGaffney嘿，我在rep0中遇到了这个错误，y-x$period-x$periodstart:invalid'times'arguments导入数据时我修改了列名，可能就是这样。粘贴并运行代码（包括数据导入）时会发生什么情况？在matrixspecialdf1中获取错误，nrow=4，ncol=7，byrow=TRUE:找不到函数special@Vaibhav现在试试，我用bigrow替换了specialdf1。我的打字错误。

m <- matrix(0, nrow=4, ncol=7)

special_spread <- function(df1){
    bigrow <- c()
    for(i in 1:nrow(df1)){
      pt1 <- rep(0, df1$SubStart[i])
      pt2 <- rep(df1$amount[i], df1$Subperiod[i])
      pt3 <- rep(0, ncol(m) - (length(pt2)+length(pt1)) )
      bigrow <- c(bigrow, pt1, pt2, pt3)
    }
    m1 <- as.data.frame(matrix(bigrow, nrow=4, ncol=7, byrow = TRUE))
    m1 <- cbind(df1$name, m1)
    colnames(m1) <- c("name", paste0("M", 0:6))
    return(m1)
}

> special_spread(df1)
  name  M0  M1  M2  M3  M4 M5 M6
1  Tom 300 300 300   0   0  0  0
2  Tom   0   0 100 100 100  0  0
3  Jim 500 500 500 500 500  0  0
4  Jim   0 600 600 600   0  0  0

library(data.table)
library(plyr)

df <- data.table(read.table(text='Name Amount Period  Start
                                   Tom  300    3      0
                                   Tom  100    3      2
                                   Jim  500    5      0
                                   Jim  600    3      1', header=T, row.names = NULL))
#Create a row by repeating df$Amount, df$Period times and padding with 0
create_rows <- function(x, y){
  c(rep(0, x$Start), rep(x$Amount, x$Period), rep(0, y - x$Period - x$Start))
}

#Create a new data.table and add Name column
df2 <- data.table(Name = df$Name)

#Create an array of month names
months <- paste('M', 0:6, sep = '')

#Use adply (from plyr) to apply create_rows() accross all rows of df2
#.expand = FALSE ensures the size of the returned data.frame is the right size
#.id = NULL stops adply() from creating an index column
#with = FALSE allows the variable month to be used to refer to columns
df2[, months := adply(df, 
                      1, create_rows,
                      length(months), 
                      .expand = FALSE, 
                      .id =NULL),
    with = FALSE]