R 有任何更优雅的解决方案可以将列添加到多个数据框中

【问题标题】：R there any more elegant solutions to add columns into multiple data framesR 有任何更优雅的解决方案可以将列添加到多个数据框中
【发布时间】：2018-08-16 16:08:53
【问题描述】：

我一直在绞尽脑汁想弄清楚如何减少所需的代码量

  • 将数据帧列表中的一列从整数更改为字符
  • 为所有数据框设置一个 data.table 键
  • 根据数据框名称向一组数据框添加一列常量

这是我的代码。我知道必须有更好的方法。

将列从整数修改为字符

MERGED1996.97.PP$OPEID<-as.character(MERGED1996.97.PP$OPEID)
MERGED1997.98.PP$OPEID<-as.character(MERGED1997.98.PP$OPEID)
MERGED1998.99.PP$OPEID<-as.character(MERGED1998.99.PP$OPEID)
MERGED1999.00.PP$OPEID<-as.character(MERGED1999.00.PP$OPEID)
MERGED2000.01.PP$OPEID<-as.character(MERGED2000.01.PP$OPEID)
MERGED2001.02.PP$OPEID<-as.character(MERGED2001.02.PP$OPEID)
MERGED2002.03.PP$OPEID<-as.character(MERGED2002.03.PP$OPEID)
MERGED2003.04.PP$OPEID<-as.character(MERGED2003.04.PP$OPEID)
MERGED2004.05.PP$OPEID<-as.character(MERGED2004.05.PP$OPEID)
MERGED2005.06.PP$OPEID<-as.character(MERGED2005.06.PP$OPEID)
MERGED2006.07.PP$OPEID<-as.character(MERGED2006.07.PP$OPEID)
MERGED2007.08.PP$OPEID<-as.character(MERGED2007.08.PP$OPEID)
MERGED2008.09.PP$OPEID<-as.character(MERGED2008.09.PP$OPEID)
MERGED2009.10.PP$OPEID<-as.character(MERGED2009.10.PP$OPEID)
MERGED2010.11.PP$OPEID<-as.character(MERGED2010.11.PP$OPEID)
MERGED2011.12.PP$OPEID<-as.character(MERGED2011.12.PP$OPEID)
MERGED2012.13.PP$OPEID<-as.character(MERGED2012.13.PP$OPEID)
MERGED2013.14.PP$OPEID<-as.character(MERGED2013.14.PP$OPEID)
MERGED2014.15.PP$OPEID<-as.character(MERGED2014.15.PP$OPEID)
MERGED2015.16.PP$OPEID<-as.character(MERGED2015.16.PP$OPEID)

设置键

setkey(MERGED199<-, "OPEas.character(
setkey(MERGED199<-, "OPEas.character(
setkey(MERGED199<-, "OPEID")
setkey(MERGED1999.00.PP, "OPEID")
setkey(MERGED2000.01.PP, "OPEID")
setkey(MERGED2001.02.PP, "OPEID")
setkey(MERGED2002.03.PP, "OPEID")
setkey(MERGED2003.04.PP, "OPEID")
setkey(MERGED2004.05.PP, "OPEID")
setkey(MERGED2005.06.PP, "OPEID")
setkey(MERGED2006.07.PP, "OPEID")
setkey(MERGED2007.08.PP, "OPEID")
setkey(MERGED2008.09.PP, "OPEID")
setkey(MERGED2009.10.PP, "OPEID")
setkey(MERGED2010.11.PP, "OPEID")
setkey(MERGED2011.12.PP, "OPEID")
setkey(MERGED2012.13.PP, "OPEID")
setkey(MERGED2013.14.PP, "OPEID")
setkey(MERGED2014.15.PP, "OPEID")
setkey(MERGED2015.16.PP, "OPEID")

添加列

MERGED1996.97.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED1996.97.PP))), 9,12)]
MERGED1997.98.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED1997.98.PP))), 9,12)]
MERGED1998.99.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED1998.99.PP))), 9,12)]
MERGED1999.00.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED1999.00.PP))), 9,12)]
MERGED2000.01.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2000.01.PP))), 9,12)]
MERGED2001.02.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2001.02.PP))), 9,12)]
MERGED2002.03.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2002.03.PP))), 9,12)]
MERGED2003.04.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2003.04.PP))), 9,12)]
MERGED2004.05.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2004.05.PP))), 9,12)]
MERGED2005.06.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2005.06.PP))), 9,12)]
MERGED2006.07.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2006.07.PP))), 9,12)]
MERGED2007.08.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2007.08.PP))), 9,12)]
MERGED2008.09.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2008.09.PP))), 9,12)]
MERGED2009.10.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2009.10.PP))), 9,12)]
MERGED2010.11.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2010.11.PP))), 9,12)]
MERGED2011.12.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2011.12.PP))), 9,12)]
MERGED2012.13.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2012.13.PP))), 9,12)]
MERGED2013.14.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2013.14.PP))), 9,12)]
MERGED2014.15.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2014.15.PP))), 9,12)]
MERGED2015.16.PP[, ay :=substr(gsub("\\.","",deparse(substitute(MERGED2015.16.PP))), 9,12)]

chinsoon12 的代码效果很好。现在我需要更改列名。我加了

colnames(df)[1] <- "UNITID"

到代码

lapply(dfLs, function(nm) {
df <- get(nm)
setDT(df)[, OPEID := as.character(OPEID)]
setkeyv(df, "OPEID")
colnames(df)[1] <- "UNITID"
df[, ay := substr(gsub("\\.", "", nm), 9, 12)]
})

但它不会修改第一列的名称。非常感谢任何建议。

谢谢。

【问题讨论】：

  • 尝试purrr::map(list_of_dataframes, ~as.character(.x$OPEID)) 将列从整数更改为字符。那应该遍历数据框并进行调整。然后可能是purrr::map(list_of_dataframes, ~setkey(.x, "OPEID")) 的键。
  • 使用dplyr::mutate_at 在位置进行变异，或dplyr::mutate_if 用于基于谓词的变异
  • 最后，purrr::map(list_of_dataframes, ~.x[, ay :=substr(gsub("\\.","",deparse(substitute(.x))), 9,12)]) 更改列。虽然没有数据，但我实际上无法对此进行测试。

标签： r dplyr data.table apply


【解决方案1】：

下面的作品：

sapply(c('data.table', 'ggplot2', 'tictoc'), library, character.only = T, quietly = T)

set.seed(28)
nobv <- 1000 # number of observations per data.table 

l <- lapply(1:5, function(z){
  ind <- sample(x = 1:nrow(diamonds), size = nobv, replace = T)
  return(as.data.table(diamonds[ind,]))
})

# Name all the data.tables
names(l) <- LETTERS[1:length(l)]

cols_to_change <- c('carat', 'z')

# changing cols_to_change to character
lapply(l, function(z){
  z[, (cols_to_change) := lapply(.SD, as.character), .SDcols = cols_to_change]
})

# set key for all data.table
key_by_cols <- c('carat', 'z')
lapply(l, function(z){
  setkeyv(z, key_by_cols)
})

# Add a name column for the data.tables 
lapply(1:length(l), function(z){
  l[[z]]$dtname <<- names(l)[z]
})

【讨论】：

    【解决方案2】：

    您可能可以这样做：

    dfLs <- c("MERGED1996.97.PP","MERGED1997.98.PP","MERGED1998.99.PP","MERGED1999.00.PP",
    "MERGED2000.01.PP","MERGED2001.02.PP","MERGED2002.03.PP","MERGED2003.04.PP",
    "MERGED2004.05.PP","MERGED2005.06.PP","MERGED2006.07.PP","MERGED2007.08.PP",
    "MERGED2008.09.PP","MERGED2009.10.PP","MERGED2010.11.PP","MERGED2011.12.PP",
    "MERGED2012.13.PP","MERGED2013.14.PP","MERGED2014.15.PP","MERGED2015.16.PP")

library(data.table)
lapply(dfLs, function(nm) {
    df <- get(nm)
    setDT(df)[, OPEID := as.character(OPEID)]
    setkeyv(df, "OPEID")
    df[, ay := substr(gsub("\\.", "", nm), 9, 12)]
})

    【讨论】：

      【解决方案3】：

      如果我理解正确，OP 必须处理包含在 20 个单独数据框中的会计数据，每个会计年度一个。

      OP 要求 减少执行某些操作所需的代码量。

      如果数据框都具有相同的结构（编号、名称、类型和列顺序），我建议将数据绑定到一个大数据对象中。这将大大简化包括分组和拆分在内的所有操作。

      library(data.table)
# create vector of dataframe names
df_names <- sapply(1996:2015, function(yr) sprintf("MERGED%4i.%02i.PP", yr, (yr + 1)%% 100))
# Alternatively, the names can be grabbed from the global environment
df_names <- grep("^MERGED.*PP$", ls(), value = TRUE)
# combine into one data object, prepend id column 
big <- rbindlist(mget(df_names), idcol = "df_name")
big

                   df_name     V1  OPEID         V3
              <char> <fctr>  <num>      <num>
 1: MERGED1996.97.PP      C 199601  0.2774292
 2: MERGED1996.97.PP      P 199602  1.0844412
 3: MERGED1997.98.PP      A 199701  0.4291247
 4: MERGED1997.98.PP      F 199702  0.5060559
 5: MERGED1998.99.PP      H 199801 -0.5466319
 6: MERGED1998.99.PP      X 199802 -0.5644520
---                                          
35: MERGED2013.14.PP      H 201301 -1.1088896
36: MERGED2013.14.PP      F 201302 -1.0149620
37: MERGED2014.15.PP      L 201401  0.5630558
38: MERGED2014.15.PP      A 201402  1.6478175
39: MERGED2015.16.PP      F 201501  1.6059096
40: MERGED2015.16.PP      W 201502 -1.1578085

      现在我们可以执行请求的操作

      # change type of OPEID column
big[, OPEID := as.character(OPEID)]
# rename second column (used to be first column before rbindlist())
setnames(BIG, 2, "UNITID")
# append column with accounting year id
big[, ay := substr(gsub("\\.", "", df_name), 9, 12)]
big[]

                   df_name UNITID  OPEID         V3     ay
              <char> <fctr>  <num>      <num> <char>
 1: MERGED1996.97.PP      C 199601  0.2774292   9697
 2: MERGED1996.97.PP      P 199602  1.0844412   9697
 3: MERGED1997.98.PP      A 199701  0.4291247   9798
 4: MERGED1997.98.PP      F 199702  0.5060559   9798
 5: MERGED1998.99.PP      H 199801 -0.5466319   9899
 6: MERGED1998.99.PP      X 199802 -0.5644520   9899
---                                                 
35: MERGED2013.14.PP      H 201301 -1.1088896   1314
36: MERGED2013.14.PP      F 201302 -1.0149620   1314
37: MERGED2014.15.PP      L 201401  0.5630558   1415
38: MERGED2014.15.PP      A 201402  1.6478175   1415
39: MERGED2015.16.PP      F 201501  1.6059096   1516
40: MERGED2015.16.PP      W 201502 -1.1578085   1516

      请注意，我没有调用 setkey()，因为由于 on = 语法和二级索引，它不再需要加入。

      可重现的数据

      警告：以下代码将在全球环境中创建 20 个数据帧。

      set.seed(1234L)
for (yr in 1996:2015) {
  nm <- sprintf("MERGED%4i.%02i.PP", yr, (yr + 1L) %% 100L)
  tmp <- data.frame(V1 = sample(LETTERS, 2L), OPEID = 100* yr + 1:2, V3 = rnorm(2L))
  assign(nm, tmp)
}

ls()

       [1] "big"              "df_names"         "MERGED1996.97.PP" "MERGED1997.98.PP" "MERGED1998.99.PP" "MERGED1999.00.PP"
 [7] "MERGED2000.01.PP" "MERGED2001.02.PP" "MERGED2002.03.PP" "MERGED2003.04.PP" "MERGED2004.05.PP" "MERGED2005.06.PP"
[13] "MERGED2006.07.PP" "MERGED2007.08.PP" "MERGED2008.09.PP" "MERGED2009.10.PP" "MERGED2010.11.PP" "MERGED2011.12.PP"
[19] "MERGED2012.13.PP" "MERGED2013.14.PP" "MERGED2014.15.PP" "MERGED2015.16.PP" "nm"               "tmp"             
[25] "yr"

      【讨论】：

        猜你喜欢
        • 2018-04-28
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        相关资源
        最近更新 更多
        热门标签
        Java Python linux javascript Mysql C# Docker 算法 前端 SpringBoot Redis Vue spring 设计模式 .net core .net kubernetes c++ 数据库 数据结构 大数据 js 机器学习 微服务 Android Go 程序员 面试 JVM ASP.net core 云原生 人工智能 后端 PHP git CSS golang k8s Nginx Django mybatis 深度学习 多线程 React 架构 devops 爬虫 云计算 Spring Boot LeetCode