嵌套循环不会带来预期的结果

Question

我对 r 还很陌生，所以这可能是一个愚蠢的疑问。

我有一个数据框，我想在其中循环浏览特定列中的行并检查其中的名称是否在另一个变量中。虽然，每一行中的值可能有多个用分号分隔的名称。我需要检查每一个。我尝试了一个嵌套循环，但我得到了重复值的列表。我的代码描述如下：

# Column in df I want to modify:
company.tickers  
----------
CARD3  
CSAN3  
CVCB3  
ELET3;ELET5;ELET6  
ENBR3  
FESA3;FESA4  
OIBR3;OIBR4  
PETR3;PETR4  
PTBL3  
TUPY3  
VLID3  

# stock names I would like to keep
stocks <- c("CARD3", "TUPY3", "OIBR3", "FESA4", "PTBL3", "VLID3","CNTO3","CSAN3","ELET3","PETR4","ENBR3")

result=list()

# Cycle through rows
for (i in 1:length(df.statements$company.tickers)){

  print(df.statements$company.tickers[i])
  stock.tickers <- strsplit(row,";")

  # Cycle through names in a cell
  for (j in 1:length(stock.tickers)) {

    if (stock.tickers[j] %in% stocks){

      print(stock.tickers[j])

      result <- c( result, stock.tickers[j])

    }

  }

}

# My expected result is the following column:
company.tickers 
----------
CARD3  
CSAN3  
CVCB3  
ELET3
ENBR3  
FESA4  
OIBR3
PETR4  
PTBL3  
TUPY3  
VLID3 

Answer 1

这里尝试使用tidyr::separate 将代码列临时拆分为单独的列。加长/整理数据，然后过滤得到你想要的。

我会感谢帮助改进这项技术的 cmets。

suppressPackageStartupMessages(library(dplyr))
suppressPackageStartupMessages(library(tidyr))
suppressPackageStartupMessages(library(purrr))

company.tickers = c(
  "CARD3",
  "CSAN3",
  "CVCB3",
  "ELET3;ELET5;ELET6",
  "ENBR3",
  "FESA3;FESA4",
  "OIBR3;OIBR4",
  "PETR3;PETR4",
  "PTBL3",
  "TUPY3",
  "VLID3")

random.data <- runif(length(company.tickers))
stocks <- c("CARD3", "TUPY3", "OIBR3", "FESA4", "PTBL3", "VLID3", "CNTO3", "CSAN3", "ELET3", "PETR4", "ENBR3")

# construct a data frame
df <- dplyr::tibble(company.tickers, random.data)

# work out how many columns `separate` will need, and create a vector of unusual column names
# it feels weird that we need to do this
# but without this I always get an error from `separate` in the next step

new_cols <- paste0("zqzcol", 1:max(map_int(strsplit(df$company.tickers, ";"), length)))

# temporarily create new columns using `separate`
# then use `pivot_longer` to reabsorb these into long, tidy data
# then filter this by what is in `stocks`
# then tidy up using `select` (optional)

df %>% 
  tidyr::separate(col = company.tickers, sep = ";", into = new_cols) %>% 
  pivot_longer(cols = starts_with("zqzcol"), values_to = "company.tickers", values_drop_na = TRUE) %>% 
  filter(company.tickers %in% stocks) %>% 
  select(company.tickers, everything(), -name)
#> Warning: Expected 3 pieces. Missing pieces filled with `NA` in 10 rows [1, 2, 3,
#> 5, 6, 7, 8, 9, 10, 11].
#> # A tibble: 10 x 2
#>    company.tickers random.data
#>    <chr>                 <dbl>
#>  1 CARD3                0.568 
#>  2 CSAN3                0.0370
#>  3 ELET3                0.119 
#>  4 ENBR3                0.276 
#>  5 FESA4                0.196 
#>  6 OIBR3                0.301 
#>  7 PETR4                0.504 
#>  8 PTBL3                0.712 
#>  9 TUPY3                0.790 
#> 10 VLID3                0.956

^{由reprex package (v0.3.0) 于 2020-03-05 创建}

Answer 2

tidyverse 替代 Wimpel 非常聪明的答案：

suppressPackageStartupMessages(library(dplyr))

company.tickers = c(
"CARD3",
"CSAN3",
"CVCB3",
"ELET3;ELET5;ELET6",
"ENBR3",
"FESA3;FESA4",
"OIBR3;OIBR4",
"PETR3;PETR4",
"PTBL3",
"TUPY3",
"VLID3")

stocks <- c("CARD3", "TUPY3", "OIBR3", "FESA4", "PTBL3", "VLID3", "CNTO3", "CSAN3", "ELET3", "PETR4", "ENBR3")

df <- dplyr::tibble(company.tickers)

filter_df <- function(x, df, col) {
  df %>% 
    dplyr::filter(stringr::str_detect(.data[[col]], x))
}

purrr::map_dfr(stocks, ~ filter_df(., df = df, col = "company.tickers")) %>% 
  dplyr::distinct()
#> # A tibble: 10 x 1
#>    company.tickers  
#>    <chr>            
#>  1 CARD3            
#>  2 TUPY3            
#>  3 OIBR3;OIBR4      
#>  4 FESA3;FESA4      
#>  5 PTBL3            
#>  6 VLID3            
#>  7 CSAN3            
#>  8 ELET3;ELET5;ELET6
#>  9 PETR3;PETR4      
#> 10 ENBR3

^{由reprex package (v0.3.0) 于 2020 年 3 月 3 日创建}

Answer 3

大概是这样的吧？

#build regex
stocks.regex <- paste0( stocks, collapse = "|")
#subset using grepl ans the new regex
subset( df, grepl( stocks.regex, df$company.tickers ) )

样本数据

library(data.table)
df <- setDF(fread("company.tickers  
CARD3  
CSAN3  
CVCB3  
ELET3;ELET5;ELET6  
ENBR3  
FESA3;FESA4  
OIBR3;OIBR4  
PETR3;PETR4  
PTBL3  
TUPY3  
VLID3", sep = ","))

stocks <- c("CARD3", "TUPY3", "OIBR3", "FESA4", "PTBL3", "VLID3","CNTO3","CSAN3","ELET3","PETR4","ENBR3")