从数据框创建转置表答案

【问题标题】：Create a transpose table from data frame从数据框创建转置表
【发布时间】：2021-03-03 19:20:45
【问题描述】：

我有一个包含以下示例数据的大数据库。我正在尝试创建一个函数，该函数可以从 Dept 列中逐一过滤 Dept，并且从过滤的子集数据中为所有过滤的经理创建一个表，如下所示。

dept 的管理者可以有很多或更少，因此它可以为过滤的管理者动态创建表。因为有时部门经理可能是 1-23 岁或更多。

我尝试帮助过滤和转置数据，然后绑定它，但这对我来说不能正常工作，因为我只精通 dplyr。

df <- data.frame(Dept = c("CA","HR","CA","HR","HR","CA","HR","HR","CA","CA","HR","CA","CA"),
                 Manager = c("AKASH","MANU","AMAN","SANU","NISH","KAMAL","VEER","SANIL","SAMEER","KANU","NUKUL","KUNAL","RAMIT"),
                 PF = c("Yes","No","Yes","Yes","Yes","No","No","Yes","No","Yes","Yes","Yes","No"),
                 Yearlybonus=c(6946,5871,0,7173,2161,3008,0,3025,4323,4196,0,5594,2313),
                 Quaterlybonus=c(2683,3846,0,2391,6716,6012,5479,3869,3764,0,4632,0,2371),
                 monthlybonus=c(4453,6466,2811,6845,4377,2617,0,7631,7761,2944,6270,3534,5856))

【问题讨论】：

标签： r function dplyr

【解决方案1】：

使用data.table 处理大型数据集

此解决方案不会将转换后的数据累积为列表。相反，它使用了一个过滤数据的函数

library('data.table')
setDT(df)
df[, `:=`(Yearlybonus = as.character(Yearlybonus),
         Quaterlybonus = as.character(Quaterlybonus), 
         monthlybonus = as.character(monthlybonus))]

这个函数是一个纯函数，因为它每次都会向函数发送一份数据（df）的副本。如果您的内存空间较少，您可以重构此函数以使用全局范围内的数据 (df)。重构后的代码看起来像注释行中的代码。

# myfun <- function(x) {
myfun <- function(df, x) {
  # wide to long
  # y <- melt( df[Dept == x], 
  df <- melt( df[Dept == x], 
             id.vars = c('Manager'), 
             measure.vars = c('PF', 'Yearlybonus', 'Quaterlybonus', 'monthlybonus'), 
             variable.name = 'T' )
  # long to wide
  # y <- dcast(y, T ~ Manager, value.var = 'value')
  df <- dcast(df, T ~ Manager, value.var = 'value')

  # add dept column
  #y[, Dept := x ]
  df[, Dept := x ]

  # set column order in memory
  #nm <- names(y)
  nm <- names(df)
  nm <- c('Dept', nm[nm != 'Dept'])
  #setcolorder(y, nm)
  setcolorder(df, nm)

  #return(y[])
  return(df[])    }
}

# create index for speed
setkey(df, Dept)

# myfun(x = 'CA')   
myfun(df = df, x = 'CA')

#  Dept             T AKASH AMAN KAMAL KANU KUNAL RAMIT SAMEER
#1:   CA            PF   Yes  Yes    No  Yes   Yes    No     No
#2:   CA   Yearlybonus  6946    0  3008 4196  5594  2313   4323
#3:   CA Quaterlybonus  2683    0  6012    0     0  2371   3764
#4:   CA  monthlybonus  4453 2811  2617 2944  3534  5856   7761

# myfun(x = 'HR')   
myfun(df = df, x = 'HR')
#   Dept             T MANU NISH NUKUL SANIL SANU VEER
#1:   HR            PF   No  Yes   Yes   Yes  Yes   No
#2:   HR   Yearlybonus 5871 2161     0  3025 7173    0
#3:   HR Quaterlybonus 3846 6716  4632  3869 2391 5479
#4:   HR  monthlybonus 6466 4377  6270  7631 6845    0

【讨论】：

【解决方案2】：

这个怎么样：

library(tidyr)
library(dplyr)
library(purrr)
map(unique(df$Dept), ~df %>% filter(Dept == .x) %>% 
  mutate(across(Yearlybonus:monthlybonus, ~as.character(.x))) %>% 
  pivot_longer(PF:monthlybonus, names_to="T", values_to="vals") %>% 
  pivot_wider(names_from="Manager", values_from="vals"))


# [[1]]
# # A tibble: 4 x 9
#   Dept  T             AKASH AMAN  KAMAL SAMEER KANU  KUNAL RAMIT
#   <chr> <chr>         <chr> <chr> <chr> <chr>  <chr> <chr> <chr>
# 1 CA    PF            Yes   Yes   No    No     Yes   Yes   No   
# 2 CA    Yearlybonus   6946  0     3008  4323   4196  5594  2313 
# 3 CA    Quaterlybonus 2683  0     6012  3764   0     0     2371 
# 4 CA    monthlybonus  4453  2811  2617  7761   2944  3534  5856 
# 
# [[2]]
# # A tibble: 4 x 8
#   Dept  T             MANU  SANU  NISH  VEER  SANIL NUKUL
#   <chr> <chr>         <chr> <chr> <chr> <chr> <chr> <chr>
# 1 HR    PF            No    Yes   Yes   No    Yes   Yes  
# 2 HR    Yearlybonus   5871  7173  2161  0     3025  0    
# 3 HR    Quaterlybonus 3846  2391  6716  5479  3869  4632 
# 4 HR    monthlybonus  6466  6845  4377  0     7631  6270

【讨论】：