有没有办法在一定数量的块之后停止 readr::read_tsv_chunked() ？

Question

我正在尝试在一个大的 .tsv 文件上使用 read_tsv_chunked()，并且想在一定数量的块之后停止。

@jimhester 提出了一种有用的方法，能够使用browse() 交互式查看给定块：https://github.com/tidyverse/readr/issues/848#issuecomment-388234659，但我想编写一个函数，1）只返回感兴趣的块；和 2) 在返回该块后停止读取文件。

我已修改 Jim 的响应以返回块，以便我可以将其与 DataFrameCallback 一起使用，但无法弄清楚如何停止从 read_tsv_chunked() 中读取。

到目前为止我的方法：

get_problem_chunk <- function(num) {
  i <- 1
  function(x, pos) {
    if (i == num) {
      i <<- i + 1
      return(x)
    }
    i <<- i + 1
    message(pos) # to see that it's scanning the whole file
    return(NULL) # break() or error() cause errors
  }
}

write_tsv(mtcars, "mtcars.tsv")
read_tsv_chunked("mtcars.tsv", DataFrameCallback$new(get_problem_chunk(3)), chunk_size = 3)

如您所见，它返回了我想要的块，但不会停止读取，直到回调不再获得任何块：

> read_tsv_chunked("mtcars.tsv", DataFrameCallback$new(get_problem_chunk(3)), chunk_size = 3)
Parsed with column specification:
cols(
  mpg = col_double(),
  cyl = col_integer(),
  disp = col_integer(),
  hp = col_integer(),
  drat = col_double(),
  wt = col_double(),
  qsec = col_double(),
  vs = col_integer(),
  am = col_integer(),
  gear = col_integer(),
  carb = col_integer()
)
1
4
<I WANT IT TO STOP HERE, BUT DON'T KNOW HOW>
10
13
16
19
22
25
28
31
# A tibble: 3 x 11
    mpg   cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb
  <dbl> <int> <int> <int> <dbl> <dbl> <dbl> <int> <int> <int> <int>
1  14.3     8   360   245  3.21  3.57  15.8     0     0     3     4
2  24.4     4    NA    62  3.69  3.19  20       1     0     4     2
3  22.8     4    NA    95  3.92  3.15  22.9     1     0     4     2

Answer 1

@jimhester 再次救援 - https://github.com/tidyverse/readr/issues/851#issuecomment-388929640

您可以使用 SideEffectCallback（这是默认设置）来执行此操作 当传递一个普通函数时）并使用

library(readr)

get_problem_chunk <- function(num) {
  i <- 1
  function(x, pos) {
    if (i == num) {
      res <<- x
      return(FALSE)
    }
    i <<- i + 1
 }
}

write_tsv(mtcars, "mtcars.tsv")
read_tsv_chunked("mtcars.tsv", get_problem_chunk(3), chunk_size = 2, col_types = cols())
#> NULL
res
#> # A tibble: 2 x 11
#>     mpg   cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb
#>   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1  18.7     8   360   175  3.15  3.44  17.0     0     0     3     2
#> 2  18.1     6   225   105  2.76  3.46  20.2     1     0     3     1

Answer 2

由于readr 包中的read_tsv_chunked() 函数没有提供停止读取的功能，我想，也许可以使用更基本的read_tsv() 函数，它确实提供了在读入后跳过和停止的可能性n 行：

require(readr)
write.table(mtcars, "mtcars.tsv", sep = "\t", quote = FALSE)

read_tsv_chunk <- function(fpath, start.row, end.row, ...) {
  # Read read_tsv() but only from row n to m
  # For the column names, read one line:
  df.1 <- suppressWarnings(read_tsv(fpath, skip = 0, n_max = 1))
  # Then read again, from start.row to end.row, both included
  skip.row = start.row - 1
  df <- suppressWarnings((read_tsv(fpath, skip = skip.row, n_max = end.row - skip.row , ...))
  colnames(df) <- colnames(df.1)
  df
}

现在：

read_tsv_chunk("mtcars.tsv", 7, 9) 
# read "mtcars.tsv" from the 7th to the 9th column (both included)

给予：

## Parsed with column specification:
## cols(
##   mpg = col_character(),
##   cyl = col_integer(),
##   disp = col_integer(),("mtcars.tsv", chunk_size=3, col_names = TRUE, skip = 6, g
##   hp = col_integer(),
##   drat = col_integer(),d("mtcars.tsv", chunk_size = 3, skip = 6, col_names = TRUE
##   wt = col_double(),
##   qsec = col_double(),
##   vs = col_double(),
##   am = col_integer(),
##   gear = col_integer(),
##   carb = col_integer()
## )
## Parsed with column specification:
## cols(
##   Valiant = col_character(),
##   `18.1` = col_double(),
##   `6` = col_integer(),
##   `225` = col_double(),
##   `105` = col_integer(),
##   `2.76` = col_double(),
##   `3.46` = col_double(),
##   `20.22` = col_double(),
##   `1` = col_integer(),
##   `0` = col_integer(),
##   `3` = col_integer(),
##   `1_1` = col_integer()
## )
## # A tibble: 3 x 12
##   mpg          cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb  `NA`
##   <chr>      <dbl> <int> <dbl> <int> <dbl> <dbl> <dbl> <int> <int> <int> <int>
## 1 Duster 360  14.3     8  360.   245  3.21  3.57  15.8     0     0     3     4
## 2 Merc 240D   24.4     4  147.    62  3.69  3.19  20.0     1     0     4     2
## 3 Merc 230    22.8     4  141.    95  3.92  3.15  22.9     1     0     4     2

`Parsed with column specification 出现两次不太好，第二次出现错误的列名...

其实你也可以这样做：

df <- read_tsv_chunked("mtcars.tsv", chunk_size = 3, skip = 6, col_names = TRUE, guess_max = 3)
df

## # A tibble: 3 x 12
##   mpg          cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb  `NA`
##   <chr>      <dbl> <int> <dbl> <int> <dbl> <dbl> <dbl> <int> <int> <int> <int>
## 1 Duster 360  14.3     8  360.   245  3.21  3.57  15.8     0     0     3     4
## 2 Merc 240D   24.4     4  147.    62  3.69  3.19  20.0     1     0     4     2
## 3 Merc 230    22.8     4  141.    95  3.92  3.15  22.9     1     0     4     2

但并不比我写的函数好... （例如，无法正确识别df$disp

如果您使用 read.table 一个非常基本的 R 函数来读取表（read.csv 和 read.delim 及其变体是 read.table 的包装函数）： 将有一个nrow = 参数来确定在多少读入行之后停止读取文件。并且有skip =参数决定开头应该跳过多少行。

read.table(file, header = TRUE, sep = "\t", quote = "\"", 
    dec = ".", fill = TRUE, comment.char = "#", nrow = 3, skip = 2 * 3)

返回你想要的：

  X18.1 X6  X225 X105 X2.76 X3.46 X20.22 X1 X0 X3 X1.1
1  14.3  8 360.0  245  3.21  3.57  15.84  0  0  3    4
2  24.4  4 146.7   62  3.69  3.19  20.00  1  0  4    2
3  22.8  4 140.8   95  3.92  3.15  22.90  1  0  4    2