This tutorial will use 4 functions within the {googledrive} package:
drive_auth()Authorize {googledrive} to access your Google Drivedrive_ls()List contents of a folder or shared drivedrive_download()Download a Drive filedrive_upload()Upload a Drive file
This code is designed to be run locally (meaning on someone’s computer manually). If the Google Drive folder is private, you need to run the authentication code first before you knit the Rmd file.
Below are examples using a NOAA Google Drive folder.
install.packages("googledrive")
I have created a Google Drive folder in my NOAA account. It is only viewable if you are in NOAA, and is only editable by the creator (Eli). It has the following sample files:
*Only people with a NOAA email will be able to see this folder
url_googledrive <- "https://drive.google.com/drive/folders/11WnXxs56jORbLkD1mFTZxwSaShex3Sse"
id_googledrive <- "11WnXxs56jORbLkD1mFTZxwSaShex3Sse"
All files.
dir_files <- googledrive::drive_ls(path = url_googledrive)
dir_files$name
## [1] "example.docx" "example2.docx" "our-google-spreadsheet"
## [4] "data.xlsx" "our-google-doc" "Stone.csv"
## [7] "Rock.csv"
All csv files.
dir_files <- googledrive::drive_ls(path = url_googledrive, type="csv")
dir_files$name
## [1] "Stone.csv" "Rock.csv"
All Excel files.
dir_files <- googledrive::drive_ls(path = url_googledrive, type="xlsx")
dir_files$name
## [1] "data.xlsx"
All Google Spreadsheets.
dir_files <- googledrive::drive_ls(path = url_googledrive, type="spreadsheet")
dir_files$name
## [1] "our-google-spreadsheet"
All Google Documents.
dir_files <- googledrive::drive_ls(path = url_googledrive, type="document")
dir_files$name
## [1] "our-google-doc"
First get the files in the folder.
a <- googledrive::drive_ls(path = url_googledrive)
Download from Google Drive and save to a folder called data.
file_name <- "data.xlsx"
# find the id of that file
file_id <- a$id[which(a$name==file_name)]
# download the file and save to data folder
googledrive::drive_download(file=file_id, overwrite = TRUE, path = file.path("data", file_name))
## File downloaded:
## • 'data.xlsx' <id: 1FL9HBcSh9QjELpiMYoR5PPEi1CIIx8g9>
## Saved locally as:
## • 'data/data.xlsx'
Read that file in using the readxl package.
tmp <- readxl::read_excel(file.path(here::here(), "data", "data.xlsx"))
tmp
## # A tibble: 6 × 2
## year count
## <dbl> <dbl>
## 1 1990 10
## 2 1991 20
## 3 1992 10
## 4 1993 30
## 5 1994 40
## 6 1995 50
Download from Google Drive and save to a folder called data.
file_name <- "Stone.csv"
file_id <- a$id[which(a$name==file_name)]
googledrive::drive_download(file=file_id, overwrite = TRUE, path = file.path("data", file_name))
## File downloaded:
## • 'Stone.csv' <id: 15nK21BniTuMpAhEXXAnsHwopOXOEcnzf>
## Saved locally as:
## • 'data/Stone.csv'
Read that file in. The first 2 lines are metadata and are skipped.
tmp <- read.csv(file.path(here::here(), "data", file_name), skip=2)
tmp
## year count
## 1 1990 10
## 2 1991 20
## 3 1992 10
## 4 1993 30
## 5 1994 40
## 6 1995 50
Download all csv files and save to a folder called data.
a <- googledrive::drive_ls(path = url_googledrive, type = "csv")
for (i in 1:nrow(a)){
googledrive::drive_download(a$id[i], overwrite = TRUE, path = file.path("data", a$name[i]))
}
## File downloaded:
## • 'Stone.csv' <id: 15nK21BniTuMpAhEXXAnsHwopOXOEcnzf>
## Saved locally as:
## • 'data/Stone.csv'
## File downloaded:
## • 'Rock.csv' <id: 1-dz_3QUqmzuXtWXfCcZUowUhZI1x9XAb>
## Saved locally as:
## • 'data/Rock.csv'
a <- googledrive::drive_ls(path = url_googledrive, type = "spreadsheet")
for (i in 1:nrow(a)){
googledrive::drive_download(a$id[i], type = "csv", overwrite = TRUE, path = file.path("data", a$name[i]))
}
## File downloaded:
## • 'our-google-spreadsheet' <id: 1plRbVAZJ_bkGFw9Y54u57eFxL-F5zaXG8elFm3r9BXs>
## Saved locally as:
## • 'data/our-google-spreadsheet.csv'
Here I show how you might use the gert package. Note this works because I am doing it within RStudio and my RStudio set-up has permission to push to this repo. If you don’t have RStudio set-up to push to GitHub, then you need to set that up first. Also you need to be working in the local repo of the GitHub repo that your are pushing to. This sounds more complicated than it is.
gert::git_add(file.path("data", "Stone.csv"))
gert::git_commit("Adding a file", author = "Eli Holmes <[email protected]>")
gert::git_push(remote = "origin", repo = ".")
You can also use these set ups to set up your code to run from GitHub via GitHub Actions. If the Google Drive is not publicly viewable, then you’ll need to research how to set up authorization in a deployed application read this. But if the Google Drive is publicly viewable, then it is fairly easy to set up a Google Action that is triggered on a schedule or triggered by an outside event. Read about how to do that here.
For this example, you will need to set authorization to allow uploading files. Run Steps 1-3 in the authorization section below with this scope.
googledrive::drive_auth(scopes = "https://www.googleapis.com/auth/drive", email = "[email protected]")
Only a person who has permission to edit the folder will be able to upload to it. See the Rmd file for this section to see how the header information is set.
Here we have an example data set of daily air quality measurements in New York, May to September 1973. We are going to save this table and share it back to Google Drive.
Create table.
airquality <- datasets::airquality
head(airquality)
## Ozone Solar.R Wind Temp Month Day
## 1 41 190 7.4 67 5 1
## 2 36 118 8.0 72 5 2
## 3 12 149 12.6 74 5 3
## 4 18 313 11.5 62 5 4
## 5 NA NA 14.3 56 5 5
## 6 28 NA 14.9 66 5 6
fil <- file.path(here::here(), "data", "airquality.csv")
write.csv(x = airquality, file = fil)
Upload table to Google Drive.
googledrive::drive_upload(media = fil, path = googledrive::as_id(id_googledrive), overwrite = TRUE)
## File trashed:
## • 'airquality.csv' <id: 104JClM466deu24G8XuycbRYjAaYCHbbp>
## Local file:
## • '/Users/eli.holmes/Documents/GitHub/RVerse-Tutorials/GoogleDrive1/data/airquality.csv'
## Uploaded into Drive file:
## • 'airquality.csv' <id: 1zOiAkKTK5kMf-vjQoeI04kIURMBWBR1V>
## With MIME type:
## • 'text/csv'
Create image.
fil <- file.path(here::here(), "data", "airquality_plot.jpeg")
jpeg(file=fil)
hist(airquality$Temp, main = "Temperature Histogram", ylab = "Temperatures")
dev.off()
## quartz_off_screen
## 2
Upload image to Google Drive.
googledrive::drive_upload(media = fil, path = googledrive::as_id(id_googledrive), overwrite = TRUE)
## Local file:
## • '/Users/eli.holmes/Documents/GitHub/RVerse-Tutorials/GoogleDrive1/data/airquality_plot.jpeg'
## Uploaded into Drive file:
## • 'airquality_plot.jpeg' <id: 1-0Hx-voNL5EmlfpZFFGHzBiI8Nbpwwli>
## With MIME type:
## • 'image/jpeg'
This Rmd will download Word (docx) files from a Google Drive folder
and convert to Rmds.
This will download the files from the folder and save to a folder called
data.
a <- googledrive::drive_ls(path = url_googledrive, type = "docx")
for (i in 1:nrow(a)){
googledrive::drive_download(a$id[i], overwrite = TRUE, path = file.path("data", a$name[i]))
}
Converting Word to Rmd works well if your Word document is simple and all the text has style of “Normal”. Click the Style pane from the Home tab in Word to see the style applied to text. Real-world Word files don’t convert so well but at least you get the text. Tables are particularly badly converted.
Note, read up on the options in Pandoc. You can tell it how to deal with track changes in the document.
for (i in 1:nrow(a)){
fil <- file.path(here::here(), "data", a$name[i])
outfil <- file.path(here::here(), "data", paste0(stringr::str_sub(a$name[i], 1, -5), "Rmd"))
rmarkdown::pandoc_convert(fil, to="markdown", output = outfil, options=c("--wrap=none", "--extract-media=."))
}
If you are just running code at the command line, then run steps 1 and 2 once. If you are running code in an Rmd, then run Steps 1-2 from the command line once and then follow the instructions in Step 3 to see what code you need to add to your Rmd file.
options(
gargle_oauth_cache = ".secrets",
gargle_oauth_email = "[email protected]"
)
This would set the authorization to readonly.
googledrive::drive_auth(scopes = "https://www.googleapis.com/auth/drive.readonly", email = "[email protected]")
This would set the authorization to full control (read, edit, upload and delete).
googledrive::drive_auth(scopes = "https://www.googleapis.com/auth/drive", email = "[email protected]")
Run that once from the command line, before knitting the Rmd. You’ll get a pop-up window asking permission for access.
Use echo=FALSE so that users don’t see this. This tells Rmd where to
find the cached token to use for authorization. Obviously, add your
email not the one here. Note, the gargle help
file
doesn’t say that this is what you should do but this is what worked for
me.
options(
gargle_oauth_cache = ".secrets",
gargle_oauth_email = "[email protected]"
)
googledrive::drive_deauth()
googledrive::drive_auth(scopes = "https://www.googleapis.com/auth/drive.readonly", email="[email protected]")
If you look at the Rmd file for this document, you’ll see this code at
the top with include=FALSE so it doesn’t show in the output. Make sure
the scopes part matches whatever you used in Step 2.
Make sure .secrets is in your .github file. Don’t push that up to
GitHub. Maybe you even want to delete that folder automatically on
signing out.