一阶差分数据帧

Question

我有以下数据框：

>dados

COUNTRY   Year   CO2 emissions Pop. Growth(%)
Argentina  1994      1.23         0.3
Argentina  1995      1.26         0.2
Argentina  1996      1.28         0.4
Argentina  1997      1.24         0.2
Brazil     1994      1.54         0.7
Brazil     1995      1.59         0.6
Brazil     1996      1.60         0.9
Brazil     1997      1.58         1.3

我想首先区分每个国家/地区的变量 CO2 emissions 和 Pop. Growth(%)。我已经尝试过 dados[,2:4] <- diff(dados[,2:4]) 函数，但它返回了错误：

“r[i1] 中的错误 - r[-length(r):-(length(r) - lag + 1L)] : 非数字 二元运算符的参数"

Answer 1

这里是dplyr:

library(dplyr)

df %>%
  group_by(COUNTRY) %>%
  mutate_at(vars(CO2_emissions:Pop_Growth), funs(.-lag(.)))

编辑：从dplyr 0.8.0 开始，funs() 已被软性弃用。对于较新版本的dplyr

，请改用以下内容

df %>%
  group_by(COUNTRY) %>%
  mutate_at(vars(CO2_emissions:Pop_Growth), list(~ .x - lag(.x)))

输出：

# A tibble: 8 x 4
# Groups:   COUNTRY [2]
  COUNTRY    Year CO2_emissions Pop_Growth
  <fct>     <int>         <dbl>      <dbl>
1 Argentina  1994         NA        NA    
2 Argentina  1995          0.03     -0.100
3 Argentina  1996          0.02      0.2  
4 Argentina  1997         -0.04     -0.2  
5 Brazil     1994         NA        NA    
6 Brazil     1995          0.05     -0.100
7 Brazil     1996          0.01      0.3  
8 Brazil     1997         -0.02      0.4 

数据：

df = structure(list(COUNTRY = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 
2L, 2L), .Label = c("Argentina", "Brazil"), class = "factor"), 
    Year = c(1994L, 1995L, 1996L, 1997L, 1994L, 1995L, 1996L, 
    1997L), CO2_emissions = c(1.23, 1.26, 1.28, 1.24, 1.54, 1.59, 
    1.6, 1.58), Pop_Growth = c(0.3, 0.2, 0.4, 0.2, 0.7, 0.6, 
    0.9, 1.3)), .Names = c("COUNTRY", "Year", "CO2_emissions", 
"Pop_Growth"), class = "data.frame", row.names = c(NA, -8L))