R - 如何导入带有块的巨大 .csv？

Question

我正在尝试导入带有块和过滤器的巨大 .csv。 但我的代码只是读取存档的一部分（2000 万 4500 万）。

我也已经尝试过使用data.table() 但没有成功。

arq_grande <- file("cnpj_dados_cadastrais_pj.csv", "r")
tam_chunk <- 5000
df1 <- read.csv(arq_grande, nrows = 10, header = T, sep = "#", dec = ".")
for(i in 1:ncol(df1)){df1[,i] <- df1[,i] %>% iconv(from = 'UTF-8', to = 'latin1')}
df_filtrado <- df1 %>% filter(codigo_natureza_juridica == c("2143","2330")) %>%  select(cnpj,everything())
write.table(df_filtrado, "/cnpj_dados_cadastrais_pj_filtrado_coop.csv", row.names = F, sep = "#", dec = ".")
names(df1)
nrow <- 1
totalRows <- 0

repeat {
  df <- read.csv(arq_grande, header=FALSE, sep="#", col.names = names(df1), nrows = tam_chunk)
  for(i in 1:ncol(df)){df[,i] <- df[,i] %>% iconv(from = 'UTF-8', to = 'latin1')}
  nRow = nrow(df)
  totalRows <- totalRows + nRow
  cat("Lendo", nrow(df), "linhas, total lido", totalRows, "\n")
  if (nrow(df) == 0)
    break

  df_filtrado <- df %>% filter(codigo_natureza_juridica == c("2143","2330")) %>%  select(cnpj,everything())
  write.table(df_filtrado, "/cnpj_dados_cadastrais_pj_filtrado_coop.csv", append = T, col.names = F, row.names = F, sep = "#", dec = ".")
}
close(arq_grande)

我在这里看到了其他示例，但没有任何效果。抱歉，我是这种数据的新手。

我只想阅读 .csv 文件的所有行。

Answer 1

您可以使用 readr::read_csv 使用 skip 和 n_max 参数以块的形式读取 csv 文件：skip 是开始时要跳过的行数，n_max 是要读取的行数之后。

library("readr")

# Example uses `#` as the separator
file <- "
lineno#X#Y#Z
1#a#b#c
2#d#e#f
3#g#h#i
4#j#k#l
5#m#n#o
6#p#q#r
7#s#t#u
8#v#w#
9#x#y#z
"

# Increase the chunk size appropriately
chunk_size <- 3

# Assumption: There is a header on the first line
# but we don't know what it is.
col_names <- TRUE
line_num <- 1

while (TRUE) {
  chunk <- read_delim(
    file, "#",
    skip = line_num,
    n_max = chunk_size,
    # On the first iteration, col_names is TRUE
    # so the first line "X,Y,Z" is assumed to be the header
    # On any subsequent iteration, col_names is a character vector
    # of the actual column names
    col_names = col_names
  )

  # If the chunk has now rows, then reached end of file
  if (!nrow(chunk)) {
    break
  }

  # Do something with the chunk of data
  print(chunk)

  # Update `col_names` so that it is equal the actual column names
  col_names <- colnames(chunk)

  # Move to the next chunk. Add 1 for the header.
  line_num <- line_num + chunk_size + (line_num == 1)
}
#> # A tibble: 3 x 4
#>   lineno X     Y     Z    
#>    <dbl> <chr> <chr> <chr>
#> 1      1 a     b     c    
#> 2      2 d     e     f    
#> 3      3 g     h     i    
#> # A tibble: 3 x 4
#>   lineno X     Y     Z    
#>    <dbl> <chr> <chr> <chr>
#> 1      4 j     k     l    
#> 2      5 m     n     o    
#> 3      6 p     q     r    
#> # A tibble: 3 x 4
#>   lineno X     Y     Z    
#>    <dbl> <chr> <chr> <chr>
#> 1      7 s     t     u    
#> 2      8 v     w     <NA> 
#> 3      9 x     y     z

^{由reprex package (v0.3.0) 于 2019 年 10 月 31 日创建}

Answer 2

您可以直接使用R中的fread函数快速加载大块.csv文件数据。

下面是一个例子：

n = 1e6
DT = data.table( a=sample(1:1000,n,replace=TRUE),
                 b=sample(1:1000,n,replace=TRUE),
                 c=rnorm(n),
                 d=sample(c("foo","bar","baz","qux","quux"),n,replace=TRUE),
                 e=rnorm(n),
                 f=sample(1:1000,n,replace=TRUE) )
DT[2,b:=NA_integer_]
DT[4,c:=NA_real_]
DT[3,d:=NA_character_]
DT[5,d:=""]
DT[2,e:=+Inf]
DT[3,e:=-Inf]

write.table(DT,"test.csv",sep=",",row.names=FALSE,quote=FALSE)
cat("File size (MB):", round(file.info("test.csv")$size/1024^2),"\n")
# 50 MB (1e6 rows x 6 columns)

system.time(DF1 <-read.csv("test.csv",stringsAsFactors=FALSE))
# 60 sec (first time in fresh R session)

system.time(DF1 <- read.csv("test.csv",stringsAsFactors=FALSE))
# 30 sec (immediate repeat is faster, varies)

system.time(DF2 <- read.table("test.csv",header=TRUE,sep=",",quote="",
    stringsAsFactors=FALSE,comment.char="",nrows=n,
    colClasses=c("integer","integer","numeric",
                 "character","numeric","integer")))

现在使用来自data.table的fread

require(data.table)

system.time(DT <- fread("test.csv"))
  #  3 sec (faster and friendlier)

当您将数据扩展到 GB 时，场景也是如此。

下面是fread函数的文档。

https://www.rdocumentation.org/packages/data.table/versions/1.12.2/topics/fread