drive is a tiny program to pull or push Google Drive files.
drive was originally developed by Burcu Dogan while working on the Google Drive team. Since she is very busy and no longer able to maintain it, I took over drive on
Thursday, 1st January 2015. This repository contains the latest version of the code.
Table of Contents
- Hyphens: - vs --
- De Initializing
- Traversal Depth
- Configuring General Settings
- Excluding And Including Objects
- Pulling And Pushing Notes
- End to End Encryption
- Sharing and Emailing
- Starring Or Unstarring
- Trashing And Untrashing
- Emptying The Trash
- Printing URL
- Editing Description
- Retrieving MD5 Checksums
- Retrieving FileId
- Retrieving Quota
- Retrieving Features
- Command Aliases
- Detecting And Fixing Clashes
- .desktop Files
- Fetching And Pruning Missing Index Files
- Drive Server
- QR Code Share
- Filing Issues
- Revoking Account Access
- Applying Patches
- Why Another Google Drive Client?
- Known Issues
- Reaching Out
go 1.9.X or higher is required. See here for installation instructions and platform installers.
- Make sure to set your GOPATH in your env, .bashrc or .bash_profile file. If you have not yet set it, you can do so like this:
cat << ! >> ~/.bashrc > export GOPATH=\$HOME/gopath > export PATH=\$GOPATH:\$GOPATH/bin:\$PATH > ! source ~/.bashrc # To reload the settings and get the newly set ones # Or open a fresh terminal
The above setup will ensure that the drive binary after compilation can be invoked from your current path.
To install from the latest source, run:
go get -u github.com/odeke-em/drive/cmd/drive
- In order to address issue #138, where debug information should be bundled with the binary, you'll need to run:
go get github.com/odeke-em/drive/drive-gen && drive-gen
go get -u github.com/odeke-em/drive/drive-google
That should produce a binary
To bundle debug information with the binary, you can run:
go get -u github.com/odeke-em/drive/drive-gen && drive-gen drive-google
- Using godep
cd $GOPATH/src/github.com/odeke-em/drive/drive-gen && godep save
- Unravelling/Restoring dependencies
cd $GOPATH/src/github.com/odeke-em/drive/drive-gen && godep restore
Please see file
drive-gen/README.md for more information.
For packages on your favorite platform, please see file Platform Packages.md.
Is your platform missing a package? Feel free to prepare / contribute an installation package and then submit a PR to add it in.
Makefile which currently supports cross compilation. Just run
make and then inspect the binaries in directory
- Supported platforms to cross compile to:
- Darwin (OS X).
Also inspect file
bin/md5Sums.txt after the cross compilation.
Optionally set the
GOOGLE_API_CLIENT_SECRET environment variables to use your own API keys.
Hyphens: - vs --
A single hyphen
- can be used to specify options. However two hyphens
-- can be used with any options in the provided examples below.
Before you can use
drive, you'll need to mount your Google Drive directory on your local file system:
drive init ~/gdrive cd ~/gdrive
Google Service Account credentials
drive init --service-account-file <gsa_json_file_path> ~/gdrive cd ~/gdrive
Where <gsa_json_file_path> must be a Google Service Account credentials file in JSON form. This feature was implemented as requested by:
The opposite of
drive init, it will remove your credentials locally as well as configuration associated files.
drive deinit [-no-prompt]
For a complete deinitialization, don't forget to revoke account access, please see revoking account access
Before talking about the features of drive, it is useful to know about "Traversal Depth".
Throughout this README the usage of the term "Traversal Depth" refers to the number of
nodes/hops/items that it takes to get from one parent to children. In the options that allow it, you'll have a flag option
-depth <n> where n is an integer
Traversal terminates on encountering a zero
A negative depth indicates infinity, so traverse as deep as you can.
A positive depth helps control the reach.
|- A/ |- B/ |- C/ |- C1 |- C2 |- C10/ |- CTX/ | - Music | - Summary.txt
Items on the first level relative to A/ ie
depth 1, we'll have:
On the third level relative to C/ ie
Items: Music, Summary.txt
The items encountered in
depth 3traversal relative to C/ are:
|- C1 |- C2 |- C10/ |- CTX/ | - Music | - Summary.txt
No items are within the reach of
depth -1relative to B/ since B/ has no children.
Items within the reach of
depth -relative to CTX/ are:
| - Music | - Summary.txt
Configuring General Settings
drive supports resource configuration files (.driverc) that you can place both globally (in your home directory) and locally(in the mounted drive dir) or in the directory that you are running an operation from, relative to the root. The entries for a .driverc file is in the form a key-value pair where the key is any of the arguments that you'd get from running
drive <command> -h # e.g drive push -h
and the value is the argument that you'd ordinarily supply on the commandline. .driverc configurations can be optionally grouped in sections. See https://github.com/odeke-em/drive/issues/778.
cat << ! >> ~/.driverc > # My global .driverc file > export=doc,pdf > depth=100 > no-prompt=true > > # For lists > [list] > depth=2 > long=true > > # For pushes > [push] > verbose=false > > # For stats > [stat] > depth=3 > > # For pulls and pushes > [pull/push] > no-clobber=true > ! cat << ! >> ~/emm.odeke-drive/.driverc > # The root main .driverc > depth=-1 > hidden=false > no-clobber=true > exports-dir=$HOME/exports > ! cat << $ >> ~/emm.odeke-drive/fall2015Classes/.driverc > # My global .driverc file > exports-dir=$HOME/Desktop/exports > export=pdf,csv,txt > hidden=true > depth=10 > exclude-ops=delete,update > $
Excluding and Including Objects
drive allows you to specify a '.driveignore' file similar to your .gitignore, in the root directory of the mounted drive. Blank lines and those prefixed by '#' are considered as comments and skipped.
cat << $ >> .driveignore > # My drive ignore file > \.gd$ > \.so$ > \.swp$ > $
Pattern matching and suffixes are done by regular expression matching so make sure to use a valid regular expression suffix.
Go doesn't have a negative lookahead mechanism ie
exclude all butwhich would normally be achieved in other languages or regex engines by "?!". See https://groups.google.com/forum/#!topic/golang-nuts/7qgSDWPIh_E. This was reported and requested in issue #535. A use case might be ignoring all but say .bashrc files or .dotfiles. To enable this, prefix "!" at the beginning of the path to achieve this behavior.
Sample .driveignore with the exclude and include clauses combined
cat << $ >> .driveignore > ^\. > !^\.bashrc # .bashrc files won't be ignored > _export$ # _export files are to be ignored > !must_export$ # the exception to the clause anything with "must_export"$ won't be ignored
pull command downloads data that does not exist locally but does remotely on Google drive, and may delete local data that is not present on Google Drive.
Run it without any arguments to pull all of the files from the current path:
To pull and decrypt your data that is stored encrypted at rest on Google Drive, use flag
See Issue #543
drive pull -decryption-password '$JiME5Umf' influx.txt
Pulling by matches is also supported
cd ~/myDrive/content/2015 drive pull -matches vines docx
To force download from paths that otherwise would be marked with no-changes
drive pull -force
To pull specific files or directories, pass in one or more paths:
drive pull photos/img001.png docs
Pulling by id is also supported
drive pull -id 0fM9rt0Yc9RTPaDdsNzg1dXVjM0E 0fM9rt0Yc9RTPaTVGc1pzODN1NjQ 0fM9rt0Yc9RTPV1NaNFp5WlV3dlU
pull optionally allows you to pull content up to a desired depth.
Say you would like to get just folder items until the second level
drive pull -depth 2 heavy-files summaries
Traverse deep to infinity and beyond
drive pull -depth -1 all-my-files
Pulling starred files is allowed as well
drive pull -starred drive pull -starred -matches content drive pull -starred -all # Pull all the starred files that aren't in the trash drive pull -starred -all -trashed # Pull all the starred files in the trash
Like most commands .driveignore can be used to filter which files to pull.
- Note: Use
drive pull -hiddento also pull files starting with
To selectively pull by type e.g file vs directory/folder, you can use flags
drive pull -files a1/b2 drive pull -directories tf1
Due to popular demand, by default, checksum verification is turned off. It was deemed to be quite vigorous and unnecessary for most cases, in which size + modTime differences are sufficient to detect file changes. The discussion stemmed from issue #117.
However, modTime differences on their own do not warrant a resync of the contents of file. Modification time changes are operations of their own and can be made:
- locally by, touching a file (chtimes).
- remotely by just changing the modTime meta data.
To turn checksum verification back on:
drive pull -ignore-checksum=false
drive also supports piping pulled content to stdout which can be accomplished by:
drive pull -piped path1 path2
- In relation to issue #529, you can change the max retry counts for exponential backoff. Using a count < 0 falls back to the default count of 20:
drive pull -retry-count 14 documents/2016/March videos/2013/September
By default, the
pull command will export Google Docs documents as PDF files. To specify other formats, use the
drive pull -export pdf,rtf,docx,txt
To explicitly export instead of using
drive pull -export pdf,rtf,docx,txt -explicitly-export
By default, the exported files will be placed in a new directory suffixed by
\_exports in the same path. To export the files to a different directory, use the
drive pull -export pdf,rtf,docx,txt -exports-dir ~/Desktop/exports
Otherwise, you can export files to the same directory as requested in issue #660,
by using pull flag
-same-exports-dir. For example:
drive pull -explicitly-export -exports-dir ~/Desktop/exp -export pdf,txt,odt -same-exports-dir Resolving... + /test-exports/few.docs + /test-exports/few + /test-exports/influx Addition count 3 Proceed with the changes? [Y/n]:y Exported '/Usersfirstname.lastname@example.org/test-exports/influx' to '/Users/emmanuelodeke/Desktop/exp/influx.pdf' Exported '/Usersemail@example.com/test-exports/influx' to '/Users/emmanuelodeke/Desktop/exp/influx.txt' Exported '/Usersfirstname.lastname@example.org/test-exports/few' to '/Users/emmanuelodeke/Desktop/exp/few.pdf' Exported '/Usersemail@example.com/test-exports/few.docs' to '/Users/emmanuelodeke/Desktop/exp/few.docs.txt' Exported '/Usersfirstname.lastname@example.org/test-exports/few.docs' to '/Users/emmanuelodeke/Desktop/exp/few.docs.odt' Exported '/Usersemail@example.com/test-exports/few.docs' to '/Users/emmanuelodeke/Desktop/exp/few.docs.pdf'
- doc, docx
- jpeg, jpg
- ppt, pptx
- txt, text
- xls, xlsx
push command uploads data to Google Drive to mirror data stored locally.
pull, you can run it without any arguments to push all of the files from the current path, or you can pass in one or more paths to push specific files or directories.
push also allows you to push content up to a desired traversal depth e.g
drive push -depth 1 head-folders
drive push path expects the path to be within the context of the drive. If the drive is locally at
drive push ~/xyz.txt may not execute as desired. Each path should reference a file or directory under the root directory of the mounted directory.
You can also push multiple paths that are children of the root of the mounted drive to a destination,
in relation to issue #612, using key
For example to push the content of
integrals/complex/compilations directly to
drive push -destination a1/b2/c3 music/Travi$+Future integrals/complex/compilations
To enable checksum verification during a push:
drive push -ignore-checksum=false
To keep your data encrypted at rest remotely on Google Drive:
drive push -encryption-password '$JiME5Umf' influx.txt
For E2E discussions, see issue #543:
drive also supports pushing content piped from stdin which can be accomplished by:
drive push -piped path
To selectively push by type e.g file vs directory/folder, you can use flags
drive push -files a1/b2 drive push -directories tf1
Like most commands .driveignore can be used to filter which files to push.
- Note: Use
drive push -hiddento also push files starting with
Here is an example using drive to backup the current working directory. It pushes a tar.gz archive created on the fly. No archive file is made on the machine running the command, so it doesn't waste disk space.
tar czf - . | drive push -piped backup-$(date +"%m-%d-%Y-"%T"").tar.gz
In response to #107 and numerous other issues related to confusion about clashing paths, drive can now auto-rename clashing files. Use flag
push, and drive will try to rename clashing files by adding a unique suffix at the end of the name, but right before the extension of a file (if the extension exists). If you haven't passed in the above
-fix-clashesflag, drive will abort on trying to deal with clashing names. If you'd like to turn off this safety, pass in flag
In relation to #57 and @rakyll's #49. A couple of scenarios in which data was getting totally clobbered and unrecoverable, drive now tries to play it safe and warn you if your data could potentially be lost e.g during a to-disk clobber for which you have no backup. At least with a push you have the luxury of untrashing content. To disable this safety, run drive with flag
drive pull -ignore-conflict collaboration_documents
Playing the safety card even more, if you want to get changes that are non clobberable ie only additions run drive with flag
drive pull -no-clobber Makefile
Ordinarily your system will not traverse nested symlinks e.g:
mkdir -p a/b mkdir -p ~/Desktop/z1/z2 && ls ~ > ~/Desktop/z1/z2/listing.txt ln -s ~/Desktop/z1/z2 a/b ls -R a # Should print only z2 and nothing inside it.
However in relation to #80, for purposes of consistency with your Drive, traversing symlinks has been added.
For safety with non clobberable changes i.e only additions:
drive push -no-clobber
- Due to the reasons above, drive should be able to warn you in case of total clobbers on data. To turn off this behaviour/safety, pass in the
drive push -force sure_of_content
To push without user input (i.e. without prompt)
drive push -quiet
drive push -no-prompt
To get Google Drive to convert a file to its native Google Docs format
drive push -convert
Extra features: to make Google Drive attempt Optical Character Recognition (OCR) for png, gif, pdf and jpg files.
drive push -ocr
Note: To use OCR, your account should have this feature. You can find out if your account has OCR allowed.
Pulling And Pushing Notes
MimeType inference is from the file's extension.
If you would like to coerce a certain mimeType that you'd prefer to assert with Google Drive pushes, use flag
-coerce-mime <short-key>See List of MIME type short keys for the full list of short keys.
drive push -coerce-mime docx my_test_doc
- Excluding certain operations can be done both for pull and push by passing in flag
drive pull -exclude-ops "delete,update" vines drive push -exclude-ops "create" sensitive_files
- To show more information during pushes or pulls e.g show the current operation,
pass in option
drive pull -verbose 2015/Photos content drive push -verbose Music Fall2014
- In relation to issue #529, you can change the max retry counts for exponential backoff. Using a count < 0 falls back to the default count of 20:
drive push -retry-count 4 a/bc/def terms
You can also specify the upload chunk size to be used to push each file, by using flag
-upload-chunk-sizewhose value is in bytes. If you don't specify this flag, by default the internal Google APIs use a value of 8MiB from constant
googleapi.DefaultUploadChunkSize. Please note that your value has to be a multiple of and atleast the minimum upload chunksize of 256KiB from constant
googleapi.MinUploadChunkSize. See https://godoc.org/google.golang.org/api/googleapi#pkg-constants. If
-upload-chunk-sizeis not set yet
-upload-chunk-sizewill be the same as
To limit the upload bandwidth, please set
-upload-rate-limit=n. It's in
nKiB/s, default is unlimited.
End to End Encryption
See Issue #543
This can be toggled when you supply a non-empty password ie
-encryption-passwordfor a push.
-decryption-passwordfor a pull.
When you supply argument
-encryption-password during a push, drive will encrypt your data
and store it remotely encrypted(stored encrypted at rest), it can only be decrypted by you when you
perform a pull with the respective arg
drive push -encryption-password '$400lsGO1Di3' few-ones.mp4 newest.mkv
drive pull -decryption-password '$400lsGO1Di3' few-ones.mp4 newest.mkv
If you supply the wrong password, you'll be warned if it cannot be decrypted
$ drive pull -decryption-password "4nG5troM" few-ones.mp4 newest.mkv message corrupt or incorrect password
To pull normally push or pull your content, without attempting any *cryption attempts, skip passing in a password and no attempts will be made.
pub command publishes a file or directory globally so that anyone can view it on the web using the link returned.
drive pub photos
- Publishing by fileId is also supported
drive pub -id 0fM9rt0Yc9RTPV1NaNFp5WlV3dlU 0fM9rt0Yc9RTPSTZEanBsamZjUXM
unpub command is the opposite of
pub. It unpublishes a previously published file or directory.
drive unpub photos
- Publishing by fileId is also supported
drive unpub -id 0fM9rt0Yc9RTPV1NaNFp5WlV3dlU 0fM9rt0Yc9RTPSTZEanBsamZjUXM
Sharing and Emailing
share command enables you to share a set of files with specific users and assign them specific roles as well as specific generic access to the files. It also allows for email notifications on share.
drive share -emails firstname.lastname@example.org,email@example.com -message "This is the substring file I told you about" -role reader,writer -type group mnt/substringfinder.c projects/kmp.c $ drive share -emails firstname.lastname@example.org,email@example.com -role reader,commenter -type user influx traversal/notes/conquest
For example to share a file with users of a mailing list and a custom message
drive share -emails firstname.lastname@example.org -message "Here is the drive code" -role group mnt/drive
- By default, an email notification is sent (even if -message is not specfified). To turn off email notification, use -notify=false
$ drive share -notify=false -emails email@example.com,firstname.lastname@example.org -role reader,commenter -type user influx traversal/notes/conquest
sharecommand also supports sharing by fileId
drive share -emails email@example.com -message "Developers, developers developers" -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
- You can also share a file to only those with the link. As per https://github.com/odeke-em/drive/issues/568, this file won't be publicly indexed. To turn this option on when sharing the file,
drive share -with-link ComedyPunchlineDrumSound.mp3
unshare command revokes access of a specific accountType to a set of files.
When no -role is given it by default assumes you want to revoke all access ie "reader", "writer", "commenter"
drive unshare -type group mnt/drive drive unshare -emails firstname.lastname@example.org,email@example.com -type user,group -role reader,commenter infinity newfiles/confidential
- Also supports unsharing by fileId
drive unshare -type group -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
Starring Or Unstarring
To star or unstar documents,
drive star information quest/A/B/C drive star -id 0fM9rt0Yc9RTPaDdsNzg1dXVjM0E 0fM9rt0Yc9RTPaTVGc1pzODN1NjQ 0fM9rt0Yc9RTPV1NaNFp5WlV3dlU
drive unstar information quest/A/B/C drive unstar -id 0fM9rt0Yc9RTPaDdsNzg1dXVjM0E 0fM9rt0Yc9RTPaTVGc1pzODN1NjQ 0fM9rt0Yc9RTPV1NaNFp5WlV3dlU
diff command compares local files with their remote equivalents. It allows for multiple paths to be passed in e.g
drive diff changeLogs.log notes sub-folders/
You can diff to a desired depth
drive diff -depth 2 sub-folders/ contacts/ listings.txt
You can also switch the base, either local or remote by using flag
drive diff -base-local=true assignments photos # To use local as the base drive diff -base-local=false infocom photos # To use remote as the base
You can only diff for short changes that is only name differences, file modTimes and types, you can use flag
drive diff -skip-content-check
Files that exist remotely can be touched i.e their modification time updated to that on the remote server using the
drive touch Photos/img001.png logs/log9907.txt
For example to touch all files that begin with digits 0 to 9:
drive touch -matches $(seq 0 9)
- Also supports touching of files by fileId
drive touch -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
- You can also touch files to a desired depth of nesting within their parent folders.
drive touch -depth 3 mnt newest flux drive touch -depth -1 -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U drive touch -depth 1 -matches $(seq 0 9)
- You can also touch and explicitly set the modification time for files by:
drive touch -time 20120202120000 ComedyPunchlineDrumSound.mp3 /share-testing/ComedyPunchlineDrumSound.mp3: 2012-02-02 12:00:00 +0000 UTC
- Specify the time format that you'd like to use when specifying the time e.g
drive touch -format "2006-01-02-15:04:05.0000Z" -time "2016-02-03-08:12:15.0070Z" outf.go /share-testing/outf.go: 2016-02-03 08:12:15 +0000 UTC
The mentioned time format has to be relative to how you would represent "Mon Jan 2 15:04:05 -0700 MST 2006". See the documentation for time formatting here time.Parse
- Specify the touch time offset from the clock on your machine where:
- minus(-) means ago e.g 30 hours ago -> -30h
- blank or plus(+) means from now e.g 10 minutes -> 10m or +10m
drive touch -duration -30h ComedyPunchlineDrumSound.mp3 outf.go /share-testing/outf.go: 2016-09-10 08:06:39 +0000 UTC /share-testing/ComedyPunchlineDrumSound.mp3: 2016-09-10 08:06:39 +0000 UTC
Trashing And Untrashing
Files can be trashed using the
drive trash Demo
To trash files that contain a prefix match e.g all files that begin with Untitled, or Make
Note: This option uses the current working directory as the parent that the paths belong to.
drive trash -matches Untitled Make
Files that have been trashed can be restored using the
drive untrash Demo
To untrash files that match a certain prefix pattern
drive untrash -matches pQueue photos Untitled
- Also supports trashing/untrashing by fileId
drive trash -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U drive untrash -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
Emptying The Trash
Emptying the trash will permanently delete all trashed files. Caution: They cannot be recovered after running this command.
Deleting items will PERMANENTLY remove the items from your drive. This operation is irreversible.
drive delete flux.mp4
drive delete -matches onyx swp
- Also supports deletion by fileIds
drive delete -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
list command shows a paginated list of files present remotely.
Run it without arguments to list all files in the current directory's remote equivalent:
Pass in a directory path to list files in that directory:
drive list photos
To list matches
drive list -matches mp4 go
-trashed option can be specified to show trashed files in the listing:
drive list -trashed photos
To get detailed information about the listings e.g owner information and the version number of all listed files:
drive list -owners -l -version
- Also supports listing by fileIds
drive list -depth 3 -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
- Listing allows for sorting by fields e.g
. To do this in reverse order, suffix_r
or-` to the selected key
e.g to first sort by modTime, then largest-to-smallest and finally most number of saves:
drive list -sort modtime,size_r,version_r Photos
- For advanced listing
drive list -skip-mime mp4,doc,txt drive list -match-mime xls,docx drive list -exact-title url_test,Photos
stat commands show detailed file information for example people with whom it is shared, their roles and accountTypes, and
fileId etc. It is useful to help determine whom and what you want to be set when performing share/unshare
drive stat mnt
stat won't recursively stat a directory, to enable recursive stating:
drive stat -r mnt
- Also supports stat-ing by fileIds
drive stat -r -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
drive stat -depth 4 -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U
The url command prints out the url of a file. It allows you to specify multiple paths relative to root or even by id
drive url Photos/2015/07/Releases intros/flux drive url -id 0Bz5qQkvRAeVEV0JtZl4zVUZFWWx 1Pwu8lzYc9RTPTEpwYjhRMnlSbDQ 0Cz5qUrvDBeX4RUFFbFZ5UXhKZm8
You can edit the description of a file like this
drive edit-desc -description "This is a new file description" freshFolders/1.txt commonCore/ drive edit-description -description "This is a new file description" freshFolders/1.txt commonCore/
Even more conveniently by piping content
cat fileDescriptions | drive edit-desc -piped targetFile influx/1.txt
Retrieving MD5 Checksums
md5sum command quickly retrieves the md5 checksums of the files on your drive. The result can be fed into the "md5sum -c" shell command to validate the integrity of the files on Drive versus the local copies.
Check that files on Drive are present and match local files:
~/MyDrive/folder$ drive md5sum | md5sum -c
Do a two-way diff (will also locate files missing on either side)
~/MyDrive/folder$ diff <(drive md5sum) <(md5sum *)
Same as above, but include subfolders
~/MyDrive/folder$ diff <(drive md5sum -r) <(find * -type f | sort | xargs md5sum)
Compare across two different Drive accounts, including subfolders
~$ diff <(drive md5sum -r MyDrive/folder) <(drive md5sum -r OtherDrive/otherfolder)
- Note: Running the 'drive md5sum' command retrieves pre-computed md5 sums from Drive; its speed is proportional to the number of files on Drive. Running the shell 'md5sum' command on local files requires reading through the files; its speed is proportional to the size of the files._
You can retrieve just the fileId for specified paths
drive id [-depth n] [paths...] drive file-id [-depth n] [paths...]
drive file-id -depth 2 dup-tests bug-reproductions # drive file-id -depth 2 dup-tests bug-reproductions FileId Relative Path "0By5qKlgRJeV2NB1OTlpmSkg8TFU" "/dup-tests" "0Bz5wQlgRJeP2QkRSenBTaUowU3c" "/dup-tests/influx_0" "0Cu5wQlgRJeV2d2VmY29HV217TFE" "/dup-tests/a" "0Cy5wQlgRJeX2WXVFMnQyQ2NDRTQ" "/dup-tests/influx" "0Cy5wQlgRJeP2YGMiOC15OEpUZnM" "/bug-reproductions" "0Cy5wQlgRJeV2MzFtTm50NVV5NW8" "/bug-reproductions/drive-406" "1xmXPziMPEgq2dK-JqaUytKz_By8S_7_RVY79ceRoZwv" "info-bulletins"
quota command prints information about your drive, such as the account type, bytes used/free, and the total amount of storage available.
features command provides information about the features present on the
drive being queried and the request limit in queries per second
drive allows you to create an empty file or folder remotely Sample usage:
drive new -folder flux drive new -mime-key doc bofx drive new -mime-key folder content drive new -mime-key presentation ProjectsPresentation drive new -mime-key sheet Hours2015Sept drive new -mime-key form taxForm2016 taxFormCounty drive new flux.txt oxen.pdf # Allow auto type resolution from the extension
The open command allows for files to be opened by the default file browser, default web browser, either by path or by id for paths that exist atleast remotely
drive open -file-browser=false -web-browser f1/f2/f3 jamaican.mp4 drive open -file-browser -id 0Bz8qQkpZAeV9T1PObvs2Y3BMQEj 0Y9jtQkpXAeV9M1PObvs4Y3BNRFk
drive allows you to copy content remotely without having to explicitly download and then reupload.
drive copy -r blobStore.py mnt flagging
drive copy blobStore.py blobStoreDuplicated.py
- Also supports copying by fileIds
drive copy -r -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U ../content
drive allows you to move content remotely between folders. To do so:
drive move photos/2015 angles library archives/storage
- Also supports moving by fileId
drive move -id 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 0fM9rt0Yc9kJRPSTFNk9kSTVvb0U ../../new_location
Google Drive supports multi-parent folder structure, where one file/folder can be placed in more than one parent folder. It consumes no extra disk space on the Cloud, but after pulling such structure it may double your files several times in your file structure. Pushing non deduplicated folder structures back may also break things, so be careful.
To place file/folder into new parent folder, keeping old one as well, use
$ drive move -keep-parent photos/2015 angles library second_parent_folder
drive allows you to rename a file/folder remotely.
Two arguments are required to rename ie
<relativePath/To/source or Id>
To perform a rename:
drive rename url_test url_test_results drive rename openSrc/2015 2015-Contributions
- Also supports renaming by fileId
drive rename 0fM9rt0Yc9RTPeHRfRHRRU0dIY97 fluxing
To turn off renaming locally or remotely, use flags
-remote=false. By default both are turned on.
drive rename -local=false -remote=true a/b/c/d/e/f flux
drive supports a few aliases to make usage familiar to the utilities in your shell e.g:
- cp : copy
- ls : list
- mv : move
- rm : delete
Detecting And Fixing Clashes
You can deal with clashes by using command
- To list clashes, you can do
drive clashes [-depth n] [paths...] drive clashes -list [-depth n] [paths...] # To be more explicit
- To fix clashes, you can do:
drive clashes -fix [-fix-mode mode] [-depth n] [paths...]
There are two available modes for
rename: this is the default behavior
trash: trashing both new and old files
As previously mentioned, Google Docs, Drawings, Presentations, Sheets etc and all files affiliated with docs.google.com cannot be downloaded raw but only exported. Due to popular demand, Linux users desire the ability to have *.desktop files that enable the file to be opened appropriately by an external opener. Thus by default on Linux, drive will create *.desktop files for files that fall into this category.
To turn off this behavior, you can set flag
-desktop-links to false e.g
drive pull -desktop-links=false
Fetching And Pruning Missing Index Files
If you would like to fetch missing index files for files that would otherwise not need any modifications, run:
drive index path1 path2 path3/path3.1 # To fetch any missing indices in those paths drive index -id 0CLu4lbUI9RTRM80k8EMoe5JQY2z
You can also fetch specific files by prefix matches
drive index -matches mp3 jpg
In case you might have deleted files remotely but never using drive, and feel like you have stale indices,
drive index -prune will search your entire indices dir for index files that do not exist remotely and remove those ones
drive index -prune
- prune-and-index To combine both operations (prune and then fetch) for indices:
drive index -all-ops
To enable services like qr-code sharing, you'll need to have the server running that will serve content once invoked in a web browser to allow for resources to be accessed on another device e.g your mobile phone
go get github.com/odeke-em/drive/drive-server && drive-server drive-server
- DRIVE_SERVER_PORT : default is 8010
- DRIVE_SERVER_HOST : default is localhost
If the above keys are not set in your env, you can do this
DRIVE_SERVER_PUB_KEY=<pub_key> DRIVE_SERVER_PRIV_KEY=<priv_key> [DRIVE...] drive-server
QR Code Share
Instead of traditionally copying long links, drive can now allow you to share a link to a file by means of a QR code that is generated after a redirect through your web browser.
From then on, you can use your mobile device or any other QR code reader to get to that file.
In order for this to run, you have to have the
As long as the server is running on a known domain, then you can start the qr-link getting ie
drive qr vines/kevin-hart.mp4 notes/caches.pdf drive qr -address http://192.168.1.113:8010 books/newest.pdf maps/infoGraphic.png drive qr -address https://my.server books/newest.pdf maps/infoGraphic.png
That should open up a browser with the QR code that when scanned will open up the desired file.
about command provides information about the program as well as that about
your Google Drive. Think of it as a hybrid between the
OR for detailed information
drive about -features -quota
help command without any arguments to see information about the commands that are available:
Pass in the name of a command to get information about that specific command and the options that can be passed to it.
drive help push
To get help for all the commands
drive help all
In case of any issue, you can file one by using command
It takes flags
-pipedis set, it expects to read the body from standard input.
A successful issue-filing request will open up the project's issue tracker in your web browser.
drive issue -title "Can't open my file" -body "Drive trips out every time" drive report-issue -title "Can't open my file" -body "Drive trips out every time" cat bugReport.txt | drive issue -piped -title "push: dump on pushing from this directory"
Revoking Account Access
To revoke OAuth Access of drive to your account, when logged in with your Google account, go to https://security.google.com/settings/security/permissions and revoke the desired permissions
drive from your computer, you'll need to take out:
- Also do not forget to revoke drive's access in case you need to uninstall it.
To apply patches of code e.g in the midst of bug fixes, you'll just need a little bit of git fiddling.
For example to patch your code with that on remote branch patch-1, you'll need to go into the source code directory, fetch all content from the git remote, checkout the patch branch then run the go installation: something like this.
cd $GOPATH/src/github.com/odeke-em/drive git fetch --all git checkout patch-1 git pull origin patch-1 go get github.com/odeke-em/drive/cmd/drive
Why Another Google Drive Client?
Background sync is not just hard, it is stupid. Here are my technical and philosophical rants about why it is not worth to implement:
Too racy. Data is shared between your remote resource, local disk and sometimes in your sync daemon's in-memory structs. Any party could touch a file at any time. It is hard to lock these actions. You end up working with multiple isolated copies of the same file and trying to determine which is the latest version that should be synced across different contexts.
It requires great scheduling to perform best with your existing environmental constraints. On the other hand, file attribute have an impact on the sync strategy. Large files block -- you wouldn't like to sit on and wait for a VM image to get synced before you can start working on a tiny text file.
It needs to read your mind to understand your priorities. Which file do you need most? It needs to read your mind to foresee your future actions. I'm editing a file, and saving the changes time to time. Why not to wait until I feel confident enough to commit the changes remotely?
drive is not a sync daemon, it provides:
Upstreaming and downstreaming. Unlike a sync command, we provide pull and push actions. The user has the opportunity to decide what to do with their local copy and when they decide to. Make some changes, either push the file remotely or revert it to the remote version. You can perform these actions with user prompt:
echo "hello" > hello.txt drive push # pushes hello.txt to Google Drive echo "more text" >> hello.txt drive pull # overwrites the local changes with the remote version
Allowing to work with a specific file or directory, optionally not recursively. If you recently uploaded a large VM image to Google Drive, yet only a few text files are required for you to work, simply only push/pull the exact files you'd like to worth with:
echo "hello" > hello.txt drive push hello.txt # pushes only the specified file drive pull path/to/a/b path2/to/c/d/e # pulls the remote directory recursively
Better I/O scheduling. One of the major goals is to provide better scheduling to improve upload/download times.
Possibility to support multiple accounts. Pull from or push to multiple Google Drive remotes. Possibility to support multiple backends. Why not to push to Dropbox or Box as well?
- It probably doesn't work on Windows.
- Google Drive allows a directory to contain files/directories with the same name. Client doesn't handle these cases yet. We don't recommend you to use
driveif you have such files/directories to avoid data loss.
- Racing conditions occur if remote is being modified while we're trying to update the file. Google Drive provides resource versioning with ETags, use Etags to avoid racy cases.
- drive rejects reading from namedPipes because they could infinitely hang. See issue #208.
Doing anything interesting with drive or want to share your favorite tips and tricks? Check out the wiki and feel free to reach out with ideas for features or requests.
This project is not supported nor maintained by Google.
Copyright 2013 Google Inc. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.