[英]R code split value by % then assign each new % value a new category
我還沒有找到一種方法來做到這一點,所以詢問是否有更簡單的方法來做到這一點。 這是數據集的示例:
Revenue Product New Code
1 223,220.00 Apple
2 386,640.40 Apple
3 19,891.95 Apple
我需要獲取每個收入行,按不同百分比分配收入,然后將每個百分比分配給一個新代碼。
舉個例子,
對於 Apple,收入應通過以下方式分配:
因此,數據集中的第一個值,Revenue =223,220.00,應該分配如下:
Revenue Product New Code
1 100,449 Apple A
2 111,610 Apple B
3 11,161 Apple C
這將增加行數。
我嘗試使用此代碼,但想知道是否有更簡單的方法可以做到這一點?
#
# libraries
#
library(dplyr)
#
# load data
#
my_data <- read.csv('sales_data_to_reclassify.csv', stringsAsFactors = FALSE)
#
# get total category revenue
#
Apple_revenue <- sum(my_data[substr(my_data$product, 1, 4) == 'Apple', 'Revenue'])
Apple_rows <- which(substr(my_data$product, 1, 4) == 'Apple')
#
# set the splits
#
splits <- list(A = 0.45,
B = 0.50,
C = 0.05)
#
# apply the splits at row level
#
for (i in Apple_rows) {
#
# revenue for this row in the original data
#
row_revenue = my_data[i, 'Revenue']
for (label in names(splits)) {
#
# grab the row
#
new_row <- my_data[i, ]
#
# calculate the revenue for this split
# and update the new row
#
new_row$Revenue <- row_revenue * splits[[label]]
#
# assign the label
#
new_row$New.Code <- label
#
# build a temporary data frame to hold the new rows
#
if (label == names(splits)[1]) {
new_rows <- new_row
} else {
new_rows <- rbind(new_rows, new_row)
}
rownames(my_data) <- NULL
Apple_rows <- which(substr(my_data$product, 1, 4) == 'Apple')
}
#
# drop the original row
#
my_data <- my_data[-i, ]
#
# add in the new rows
#
my_data <- rbind(my_data, new_rows)
}
#
# test revenue
#
Apple_new_revenue <- sum(my_data[substr(my_data$product, 1, 4) == 'Apple', 'Revenue'])
這是一個非常簡單的dplyr
解決方案:
df %>%
filter(Product %in% c("Apple", "Microsoft", "Samsung") %>%
mutate(A = Revenue * 0.45,
B = Revenue * 0.50,
C = Revenue * 0.05) %>%
select(-Revenue) %>%
pivot_longer(-Product, values_to = "Revenue") %>%
rename(`New Code` = name) %>%
select(Revenue, Product, `New Code`)
這給了我們:
Revenue Product `New Code`
<dbl> <chr> <chr>
1 100449 Apple A
2 111610 Apple B
3 11161 Apple C
4 173988. Apple A
5 193320. Apple B
6 19332. Apple C
7 8951. Apple A
8 9946. Apple B
9 995. Apple C
這是一個更長但類似的base R
解決方案:
# Remove commas from Revenue and convert to numeric
df$Revenue <- as.numeric(gsub(",", "", df$Revenue))
df <- subset(df, df$Product %in% c("Apple", "Microsoft", "Samsung"))
# Calculate percentage distributions
df$A <- df$Revenue * 0.45
df$B <- df$Revenue * 0.50
df$C <- df$Revenue * 0.05
# Reshape data to long
df <- reshape(df,
varying = c("A","B","C"),
v.names = "Revenue",
direction = "long")
# Sort by ID and recode values
df <- df[order(df$id),]
df$time[df$time == 1] <- "A"
df$time[df$time == 2] <- "B"
df$time[df$time == 3] <- "C"
# Drop ID column
df <- subset(df, select = -c(id))
# Rename 'time' to 'New Code'
names(df)[3] <- "New Code"
這給了我們:
Revenue Product New Code
1: 100449.0000 Apple A
2: 111610.0000 Apple B
3: 11161.0000 Apple C
4: 173988.1800 Apple A
5: 193320.2000 Apple B
6: 19332.0200 Apple C
7: 8951.3775 Apple A
8: 9945.9750 Apple B
9: 994.5975 Apple C
使用合並:
# example data - updated to include more products.
my_data <- read.table(text = "Revenue Product
223220.0 Apple
386640.4 Apple
19891.95 Pear", header = TRUE)
# define shares
splits <- data.frame(Product = rep(c("Apple", "Pear"), c(3, 3)),
NewCode = c("A", "B", "C", "X", "Y", "Z"),
Share = c(0.45, 0.50, 0.05, 0.2, 0.3, 0.5))
# merge and get revenue shares
res <- merge(my_data, splits, by = "Product")
res$RevenueShare <- res$Revenue * res$Share
res
# Product Revenue NewCode Share RevenueShare
# 1 Apple 223220.00 A 0.45 100449.000
# 2 Apple 223220.00 B 0.50 111610.000
# 3 Apple 223220.00 C 0.05 11161.000
# 4 Apple 386640.40 A 0.45 173988.180
# 5 Apple 386640.40 B 0.50 193320.200
# 6 Apple 386640.40 C 0.05 19332.020
# 7 Pear 19891.95 X 0.20 3978.390
# 8 Pear 19891.95 Y 0.30 5967.585
# 9 Pear 19891.95 Z 0.50 9945.975
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.