ffsave: Save R and ff objects

Description

ffsave writes an external representation of R and ff objects to an ffarchive. The objects can be read back from the file at a later date by using the function ffload.

Usage

ffsave(...
, list = character(0L)
, file = stop("'file' must be specified")
, envir = parent.frame()
, rootpath = NULL
, add = FALSE

, move = FALSE
, compress = !move
, compression_level = 6
, precheck=TRUE
)
ffsave.image(file = stop("'file' must be specified"), safe = TRUE, ...)

Arguments

…

For ffsave the names of the objects to be saved (as symbols or character strings), for ffsave.image further arguments passed to ffsave

list

A character vector containing the names of objects to be saved.

file

A name for the the ffarchive, i.e. the two files <file>.RData and <file>.ffData

envir

environment to search for objects to be saved.

add

logical indicating whether the objects shall be added to the ffarchive (in this case rootpath is taken from an existing archive)

move

logical indicating whether ff files shall be moved instead of copied into the <file>.ffData

compress

logical specifying whether saving to a named file is to use compression.

compression_level

compression level passed to zip, default 6

rootpath

optional path component that all all ff files share and that can be dropped/replaced when calling ffload

precheck

logical: should the existence of the objects be checked before starting to save (and in particular before opening the file/connection)?

safe

logical. If TRUE, a temporary file is used for creating the saved workspace. The temporary file is renamed to <file>.ffData if the save succeeds. This preserves an existing workspace <file>.ffData if the save fails, but at the cost of using extra disk space during the save.

Value

a character vector with messages returned from the zip utility (one for each ff file zipped)

Details

ffsave stores objects and ff files in an ffarchive named <file>: i.e. it saves all specified objects via save in a file named <file>.RData and saves all ff files related to these objects in a zipfile named <file>.ffData using an external zip utility.

By default files are stored relative to the rootpath="\"} and will be restored relative to \code{"\" (in its original location). By providing a partial path prefix via argument rootpath the files are stored relative to this rootpath. The rootpath is stored in the <file>.RData with the name .ff.rootpath. I.e. even if the ff objects were saved with argument rootpath to ffsave, ffload by default restores in the original location. By using argument rootpath to ffload you can restore relative to a different rootpath (and using argument rootpath to ffsave gave you shorter relative paths)

By using argument add in ffsave you can add more objects to an existing ffarchive and by using argument list in ffload you can selectively restore objects.

The content of the ffarchive can be inspected using ffinfo before actually loading any of the objects.

The ffarchive can be deleted from disk using ffdrop.

Examples

Run this code

# NOT RUN {
  
# }
# NOT RUN {
  message("let's create some ff objects")
  n <- 8e3
  a <- ff(sample(n, n, TRUE), vmode="integer", length=n, filename="d:/tmp/a.ff")
  b <- ff(sample(255, n, TRUE), vmode="ubyte", length=n, filename="d:/tmp/b.ff")
  x <- ff(sample(255, n, TRUE), vmode="ubyte", length=n, filename="d:/tmp/x.ff")
  y <- ff(sample(255, n, TRUE), vmode="ubyte", length=n, filename="d:/tmp/y.ff")
  z <- ff(sample(255, n, TRUE), vmode="ubyte", length=n, filename="d:/tmp/z.ff")
  df <- ffdf(x=x, y=y, z=z)
  rm(x,y,z)

  message("save all of them")
  ffsave.image("d:/tmp/x")
  str(ffinfo("d:/tmp/x"))

  message("save some of them with shorter relative pathnames ...")
  ffsave(a, b, file="d:/tmp/y", rootpath="d:/tmp")
  str(ffinfo("d:/tmp/y"))

  message("... and add others later")
  ffsave(df, add=TRUE, file="d:/tmp/y", rootpath="d:/tmp")
  str(ffinfo("d:/tmp/y"))

  message("... and add others later")
  system.time(ffsave(a, file="d:/tmp/z", move=TRUE))
  ffinfo("d:/tmp/z")

  message("let's delete/close/remove all objects")
  close(a)  # no file anymore, since we moved a into the ffarchive
  delete(b, df)
  rm(df, a, b, n)
  message("prove it")
  ls()

  message("restore all but ff files in a different directory")
  system.time(ffload("d:/tmp/x", rootpath="d:/tmp2"))
  lapply(ls(), function(i)filename(get(i)))

  delete(a, b, df)
  rm(df, a, b)

  ffdrop(c("d:/tmp/x", "d:/tmp/y", "d:/tmp/z"))
  
# }

Run the code above in your browser using DataLab

Data Engineering and BI courses are free this week!