在文件夹中查找数据并提供有关丢失数据的反馈

Find data in folders and give feedback about missing data

我有一个 R 脚本来创建一个大数据集(实际上是一个欧洲数据集)的几个小部分。与使用一个大数据集相比,我们需要这些小部分(图块)来更轻松地编辑这些图块。

现在我有 1 个 windows 文件夹,在这个文件夹中我有 966 个自动生成的文件夹 - 每个文件夹有 4 个数据集(我希望至少是 4 个)。我们需要知道文件夹中是否恰好有这 4 个数据集,如果缺少某些数据集,我们需要知道是哪一个。您可以在下面看到的代码正在创建文件夹。它的发布只是为了让您了解结构。

in_file <- "P:/High_Resolution_Layers/Forest... .tif/2015/TCD_2015_020m_eu_03035_d04_full/TCD_2015_020m_eu_03035_d04_full.tif"


for (t in 1:length(tiles)){

  tileID <- tiles[t]

  out_dir <- file.path(output_dir,tileID)
  # out_dir_tmp <- file.path(out_dir, "tmp")
  if(!exists(out_dir)) {dir.create(out_dir, recursive = T)}
  # if(!exists(out_dir)) {dir.create(out_dir_tmp, recursive = T)}

  # tmp_file <- file.path(out_dir_tmp, paste0(tileID, "_HRL_Forest.tif")) ## <- ändern ("_HRL_Forest.tif", _clc_2012.tif, _clc_2018.tif, _slope.tif)
  out_file <- file.path(out_dir, paste0(tileID, "_HRL_Forest.tif")) ## <- ändern ("_HRL_Forest.tif", _clc_2012.tif, _clc_2018.tif, _slope.tif)

  cmd <- paste("gdalwarp",
               "-overwrite",
               "-s_srs EPSG:3035",
               "-t_srs EPSG:3035",
               "-r near",
               "-q",
               "-tr 20 20",
               "-te ", tile_list[t,3],tile_list[t,4],tile_list[t,3]+100000, tile_list[t,4]+100000,
               "-tap",
               "-of GTiff",
               in_file,
               out_file)

  system(osgeo, input=cmd)

  # cmd <- sprintf('gdal_translate -ot Byte -a_nodata 255 -co "COMPRESS=LZW" %s %s', tmp_file, out_file)
  # system(osgeo, input=cmd)

  # unlink(out_dir_tmp,recursive=T)
}

我要组成一个结构和文件列表。

  • 目录 AD
  • 每个目录必须有文件 a.tif 虽然 c.tif

由于所有目录中必须包含相同的文件,我们可以对它们进行笛卡尔/outer 连接:

dirs <- LETTERS[1:4]
files_each_dir <- paste0(letters[1:3], ".tif")
(all_files <- outer(dirs, files_each_dir, file.path))
#      [,1]      [,2]      [,3]     
# [1,] "A/a.tif" "A/b.tif" "A/c.tif"
# [2,] "B/a.tif" "B/b.tif" "B/c.tif"
# [3,] "C/a.tif" "C/b.tif" "C/c.tif"
# [4,] "D/a.tif" "D/b.tif" "D/c.tif"

由于我们不需要 matrix,我将取消列出它们,然后创建 dirs/files:

c(all_files)
#  [1] "A/a.tif" "B/a.tif" "C/a.tif" "D/a.tif" "A/b.tif" "B/b.tif" "C/b.tif"
#  [8] "D/b.tif" "A/c.tif" "B/c.tif" "C/c.tif" "D/c.tif"
for (d in dirs) dir.create(d)
for (p in all_files) writeLines(p, p)

所有预期的文件都存在

(files_found <- list.files(pattern = "*.tif", recursive = TRUE, full.names = TRUE))
#  [1] "./A/a.tif" "./A/b.tif" "./A/c.tif" "./B/a.tif" "./B/b.tif" "./B/c.tif"
#  [7] "./C/a.tif" "./C/b.tif" "./C/c.tif" "./D/a.tif" "./D/b.tif" "./D/c.tif"
### remove the leading "./"
(files_found <- gsub("^\./", "", files_found))
#  [1] "A/a.tif" "A/b.tif" "A/c.tif" "B/a.tif" "B/b.tif" "B/c.tif" "C/a.tif"
#  [8] "C/b.tif" "C/c.tif" "D/a.tif" "D/b.tif" "D/c.tif"
all(all_files %in% files_found)
# [1] TRUE
all_files[!all_files %in% files_found]
# character(0)

测试丢失的文件:

file.remove("B/c.tif")
# [1] TRUE
files_found <- list.files(pattern = "*.tif", recursive = TRUE, full.names = TRUE)
files_found <- gsub("^\./", "", files_found)
all_files[!all_files %in% files_found]
# [1] "B/c.tif"

注意:我们不使用 files_each_dir 进行任何后续测试。只有当我们需要一组固定的文件名时才需要它。

计算每个目录中的文件数

如果文件名可能不同,那么我们可以统计每个目录中的文件数,而不考虑实际名称。

(len3 <- lengths(split(files_found, sapply(strsplit(files_found, "[/\]"), `[[`, 1))) == 3)
#     A     B     C     D 
#  TRUE FALSE  TRUE  TRUE 
names(len3)[ !len3 ]
# [1] "B"

文件内容

如果您需要测试内容以使某些条件为真,请尝试这样的操作。在这里,我使用简单的 shell 命令 grep,但是任何函数(R 或 shell)需要一个路径和 returns 你需要的东西(大小,属性, 等) 应该工作。

func <- function(path) length(system2("grep", c("-lE", "'[a-z]'", path), stdout = TRUE)) > 0
(proper_contents <- sapply(files_found, func))
# A/a.tif A/b.tif A/c.tif B/a.tif B/b.tif C/a.tif C/b.tif C/c.tif D/a.tif D/b.tif 
#    TRUE    TRUE    TRUE    TRUE    TRUE    TRUE    TRUE    TRUE    TRUE    TRUE 
# D/c.tif 
#    TRUE 

让我们更改一个文件的内容进行测试:

writeLines("123", "D/a.tif")
proper_contents <- sapply(files_found, func)
# Warning in system2("grep", c("-lE", "'[a-z]'", path), stdout = TRUE) :
#   running command '"grep" -lE '[a-z]' D/a.tif' had status 1
names(proper_contents)[ !proper_contents ]
# [1] "D/a.tif"