Package {zip}

Title: Cross-Platform 'zip' Compression
Version: 3.0.0
Description: Cross-Platform 'zip' Compression Library. A replacement for the 'zip' function, that does not require any additional external tools on any platform.
License: MIT + file LICENSE
URL: https://github.com/r-lib/zip, https://r-lib.github.io/zip/
BugReports: https://github.com/r-lib/zip/issues
LinkingTo: cli
Suggests: callr, cli, curl, pillar, processx, R6, testthat, webfakes, withr
Config/Needs/website: tidyverse/tidytemplate
Config/testthat/edition: 3
Config/testthat/parallel: true
Config/testthat/start-first: large-files, http
Config/usethis/last-upkeep: 2025-05-07
Encoding: UTF-8
Config/roxygen2/version: 8.0.0
NeedsCompilation: yes
Packaged: 2026-06-10 11:46:09 UTC; gaborcsardi
Author: Gábor Csárdi [aut, cre], Kuba Podgórski [ctb], Rich Geldreich [ctb], Arm Limited [ctb, cph] (bundled Mbed TLS crypto subset in src/mbedtls/ (Apache-2.0)), Posit Software, PBC ROR ID [cph, fnd]
Maintainer: Gábor Csárdi <csardi.gabor@gmail.com>
Repository: CRAN
Date/Publication: 2026-06-10 12:40:02 UTC

zip: Cross-Platform 'zip' Compression

Description

Cross-Platform 'zip' Compression Library. A replacement for the 'zip' function, that does not require any additional external tools on any platform.

Author(s)

Maintainer: Gábor Csárdi csardi.gabor@gmail.com

Authors:

Other contributors:

See Also

Useful links:

Compress a raw GZIP stream

Description

Compress a raw GZIP stream

Usage

deflate(buffer, level = 6L, pos = 1L, size = NULL)

Arguments

buffer

Raw vector, containing the data to compress.

level

Compression level, integer between 1 (fatest) and 9 (best).

pos

Start position of data to compress in buffer.

size

Compressed size estimate, or NULL. If not given, or too small, the output buffer is resized multiple times.

Value

Named list with three entries:

See Also

base::memCompress() does the same with type = "gzip", but it does not tell you the number of bytes read from the input.

Examples

data_gz <- deflate(charToRaw("Hello world!"))
inflate(data_gz$output)

Uncompress a raw GZIP stream

Description

Uncompress a raw GZIP stream

Usage

inflate(buffer, pos = 1L, size = NULL, raw = FALSE)

Arguments

buffer

Raw vector, containing the data to uncompress.

pos

Start position of data to uncompress in buffer.

size

Uncompressed size estimate, or NULL. If not given, or too small, the output buffer is resized multiple times.

raw

Whether buffer contains a raw DEFLATE stream, i.e. one without a zlib header and trailer. The default (FALSE) expects a zlib stream.

Value

Named list with three entries:

See Also

base::memDecompress() does the same with type = "gzip", but it does not tell you the number of bytes read from the input.

Examples

data_gz <- deflate(charToRaw("Hello world!"))
inflate(data_gz$output)

Uncompress 'zip' Archives

Description

unzip() always restores modification times of the extracted files and directories.

Usage

unzip(
  zipfile,
  files = NULL,
  overwrite = TRUE,
  junkpaths = FALSE,
  exdir = ".",
  encoding = NULL,
  password = NULL
)

Arguments

zipfile

Path to the zip file to uncompress, or a character vector of paths. When multiple paths are given and all other arguments are at their defaults, the files are unzipped concurrently in a thread pool. Set the zip_threads option or the ZIP_THREADS environment variable to control the number of threads used. By default zip uses two threads.

files

Character vector of files to extract from the archive. Files within directories can be specified, but they must use a forward slash as path separator, as this is what zip files use internally. If NULL, all files will be extracted.

overwrite

Whether to overwrite existing files. If FALSE and a file already exists, then an error is thrown.

junkpaths

Whether to ignore all directory paths when creating files. If TRUE, all files will be created in exdir.

exdir

Directory to uncompress the archive to. If it does not exist, it will be created.

encoding

Encoding to use for entry filenames. ZIP files signal UTF-8 filenames via a flag in each entry; those are always decoded as UTF-8 regardless of encoding. For entries without that flag, encoding is used; NULL (the default) falls back to IBM CP437, which is what the ZIP specification prescribes for legacy entries. The value is passed to iconv().

password

Password for decrypting encrypted entries. It can be a string, a raw vector, or a function that returns one of these. If NULL (the default), the zip_password option is used, or no password if that is also NULL. The password is silently ignored for entries that are not encrypted.

Value

A data frame with one row per extracted entry and columns, invisibly: filename (path within the archive), compressed_size, uncompressed_size, timestamp, permissions, crc32, offset, type (same as in zip_list()), and path (absolute path to the extracted file on disk).

Permissions

If the zip archive stores permissions and was created on Unix, the permissions will be restored.

See Also

Other zip/unzip functions: zip_list()

Examples

## temporary directory, to avoid messing up the user's workspace.
dir.create(tmp <- tempfile())
dir.create(file.path(tmp, "mydir"))
cat("first file", file = file.path(tmp, "mydir", "file1"))
cat("second file", file = file.path(tmp, "mydir", "file2"))

zipfile <- tempfile(fileext = ".zip")
zip::zip(zipfile, "mydir", root = tmp)

## List contents
zip_list(zipfile)

## Extract and inspect result
tmp2 <- tempfile()
result <- unzip(zipfile, exdir = tmp2)
result[, c("filename", "path")]

Class for an external unzip process

Description

unzip_process() returns an R6 class that represents an unzip process. It is implemented as a subclass of processx::process.

Usage

unzip_process()

Value

An unzip_process R6 class object, a subclass of processx::process, or a subclass of callr::r_process when the fallback is active (see the Fallback section below).

Using the unzip_process class

up <- unzip_process()$new(zipfile, exdir = ".", poll_connection = TRUE,
                           stderr = tempfile(), ...)

See processx::process for the class methods.

Arguments:

Fallback

unzip_process() normally runs the bundled cmdunzip native executable via processx::process. If the executable cannot be found or fails its self-test it falls back to running unzip() in a background R process via callr::r_process. This may happen when system policies do not allow starting the cmdunzip executable., The fallback class has the same interface but inherits from callr::r_process instead of processx::process.

Set the environment variable R_ZIP_PROCESS_FALLBACK=true to force the fallback unconditionally.

Encoding

The unzip_process class does not support the encoding argument of unzip(). Non-UTF-8 filenames are decoded using the IBM CP437 fallback. Use unzip() directly if you need to handle ZIP files with filenames in other encodings (e.g. CP932).

Examples

ex <- system.file("example.zip", package = "zip")
tmp <- tempfile()
up <- unzip_process()$new(ex, exdir = tmp)
up$wait()
up$get_exit_status()
dir(tmp)

Compress Files into 'zip' Archives

Description

zip() creates a new zip archive file.

Usage

zip(
  zipfile,
  files,
  recurse = TRUE,
  compression_level = 9,
  include_directories = TRUE,
  root = ".",
  mode = c("mirror", "cherry-pick"),
  keys = NULL,
  password = NULL,
  encryption = c("aes256", "aes128", "zipcrypto")
)

zipr(
  zipfile,
  files,
  recurse = TRUE,
  compression_level = 9,
  include_directories = TRUE,
  root = ".",
  mode = c("cherry-pick", "mirror"),
  keys = NULL,
  password = NULL,
  encryption = c("aes256", "aes128", "zipcrypto")
)

zip_append(
  zipfile,
  files,
  recurse = TRUE,
  compression_level = 9,
  include_directories = TRUE,
  root = ".",
  mode = c("mirror", "cherry-pick"),
  keys = NULL,
  password = NULL,
  encryption = c("aes256", "aes128", "zipcrypto")
)

zipr_append(
  zipfile,
  files,
  recurse = TRUE,
  compression_level = 9,
  include_directories = TRUE,
  root = ".",
  mode = c("cherry-pick", "mirror"),
  keys = NULL,
  password = NULL,
  encryption = c("aes256", "aes128", "zipcrypto")
)

Arguments

zipfile

The zip file to create. If the file exists, zip overwrites it, but zip_append appends to it. If it is a directory an error is thrown.

files

Character vector of paths to files to add to the archive. See details below about absolute and relative path names.

recurse

Whether to add the contents of directories recursively.

compression_level

A number between 1 and 9. 9 compresses best, but it also takes the longest.

include_directories

Whether to explicitly include directories in the archive. Including directories might confuse MS Office when reading docx files, so set this to FALSE for creating them.

root

Change to this working directory before creating the archive.

mode

Selects how files and directories are stored in the archive. It can be "mirror" or "cherry-pick". See "Relative Paths" below for details.

keys

An optional character vector of the same length as files, specifying the paths of the corresponding entries inside the zip archive. For a file, the key is the exact archive path. For a directory, the key becomes the directory prefix under which all contents are stored. If NULL (default), paths are determined by mode. "." may not appear in files when keys is specified.

password

Password for encrypting the archive entries. It can be a string, a raw vector of bytes, or a zero-argument function that returns one of these. If NULL (the default), the zip_password option is consulted; if that is also NULL, entries are stored unencrypted. The password is interpreted as UTF-8 bytes regardless of the current locale, which matches the WinZip/7-Zip convention and ensures interoperability across platforms.

encryption

Encryption scheme to use when password is not NULL. "aes256" (the default) and "aes128" use WinZip AES encryption (AES-256 or AES-128 in CTR mode, key derived via PBKDF2-HMAC-SHA1, with an HMAC-SHA1 authentication tag). This scheme is supported by 7-Zip, WinZip, and macOS Archive Utility. "zipcrypto" uses the legacy PKWARE ZipCrypto stream cipher, which is cryptographically weak and should only be used for compatibility with tools that do not support AES encryption.

Details

zip_append() appends compressed files to an existing 'zip' file.

Relative paths

zip() and zip_append() can run in two different modes: mirror mode and cherry picking mode. They handle the specified files differently.

Mirror mode

Mirror mode is for creating the zip archive of a directory structure, exactly as it is on the disk. The current working directory will be the root of the archive, and the paths will be fully kept. zip changes the current directory to root before creating the archive.

E.g. consider the following directory structure: 

.
|-- foo
|   |-- bar
|   |   |-- file1
|   |   `-- file2
|   `-- bar2
`-- foo2
    `-- file3

Assuming the current working directory is foo, the following zip entries are created by zip: 

setwd("foo")
zip::zip("../test.zip", c("bar/file1", "bar2", "../foo2"))
#> Warning in warn_for_dotdot(data$key): Some paths reference parent directory,
#> creating non-portable zip file
zip_list("../test.zip")[, "filename", drop = FALSE]
#> # A data frame: 4 x 1
#>   filename     
#>   <chr>        
#> 1 bar/file1    
#> 2 bar2/        
#> 3 ../foo2/     
#> 4 ../foo2/file3

Note that zip refuses to store files with absolute paths, and chops off the leading / character from these file names. This is because only relative paths are allowed in zip files.

Cherry picking mode

In cherry picking mode, the selected files and directories will be at the root of the archive. This mode is handy if you want to select a subset of files and directories, possibly from different paths and put all of them in the archive, at the top level.

Here is an example with the same directory structure as above: 

zip::zip(
  "../test2.zip",
  c("bar/file1", "bar2", "../foo2"),
  mode = "cherry-pick"
)
zip_list("../test2.zip")[, "filename", drop = FALSE]
#> # A data frame: 4 x 1
#>   filename  
#>   <chr>     
#> 1 file1     
#> 2 bar2/     
#> 3 foo2/     
#> 4 foo2/file3

From zip version 2.3.0, "." has a special meaning in the files argument: it will include the files (and possibly directories) within the current working directory, but not the working directory itself. Note that this only applies to cherry picking mode.

Permissions:

zip() (and zip_append(), etc.) add the permissions of the archived files and directories to the ZIP archive, on Unix systems. Most zip and unzip implementations support these, so they will be recovered after extracting the archive.

Note, however that the owner and group (uid and gid) are currently omitted, even on Unix.

zipr() and zipr_append()

These functions exist for historical reasons. They are identical to zip() and zip_append() with a different default for the mode argument.

Value

The name of the created zip file, invisibly.

Examples

## Some files to zip up. We will run all this in the R session's
## temporary directory, to avoid messing up the user's workspace.
dir.create(tmp <- tempfile())
dir.create(file.path(tmp, "mydir"))
cat("first file", file = file.path(tmp, "mydir", "file1"))
cat("second file", file = file.path(tmp, "mydir", "file2"))

zipfile <- tempfile(fileext = ".zip")
zip::zip(zipfile, "mydir", root = tmp)

## List contents
zip_list(zipfile)

## Add another file
cat("third file", file = file.path(tmp, "mydir", "file3"))
zip_append(zipfile, file.path("mydir", "file3"), root = tmp)
zip_list(zipfile)

List Files in a 'zip' Archive

Description

List Files in a 'zip' Archive

Usage

zip_list(zipfile, encoding = NULL)

Arguments

zipfile

Path to an existing ZIP file.

encoding

Encoding to use for entry filenames. ZIP files signal UTF-8 filenames via a flag in each entry; those are always decoded as UTF-8 regardless of encoding. For entries without that flag, encoding is used; NULL (the default) falls back to IBM CP437, which is what the ZIP specification prescribes for legacy entries. The value is passed to iconv().

Details

Note that crc32 is formatted using as.hexmode(). offset refers to the start of the local zip header for each entry. Following the approach of seek() it is stored as a numeric rather than an integer vector and can therefore represent values up to 2^53-1 (9 PB).

Value

A data frame with columns: filename, compressed_size, uncompressed_size, timestamp, permissions, crc32, offset, type and encryption. type is one of file, block_device, character_device, directory, FIFO, symlink or socket. encryption is one of none, aes128, aes192, aes256, zipcrypto, or NA if encrypted but the scheme cannot be determined.

See Also

Other zip/unzip functions: unzip()

Class for an external zip process

Description

zip_process() returns an R6 class that represents a zip process. It is implemented as a subclass of processx::process.

Usage

zip_process()

Value

A zip_process R6 class object, a subclass of processx::process.

Using the zip_process class

zp <- zip_process()$new(zipfile, files, recurse = TRUE,
                         poll_connection = TRUE,
                         stderr = tempfile(), ...)

See processx::process for the class methods.

Arguments:

Examples

dir.create(tmp <- tempfile())
write.table(iris, file = file.path(tmp, "iris.ssv"))
zipfile <- tempfile(fileext = ".zip")
zp <- zip_process()$new(zipfile, tmp)
zp$wait()
zp$get_exit_status()
zip_list(zipfile)