Error in value[[3L]](cond) : No region provided (Package noctua)

Question

I am making a shinydashboard app in R. I wanted to make an app that withdraws data from a database in AWS Athena every time it is deployed and display a DT table in my shiny app that gives you the option to see the information based on what date ranges you select in dateInputRange.

I needed an R package for this so I used RAthena I also used the packages DBI for the AWS CLI and paws. I downloaded all what I needed for this like boto3 and so forth. I also made a aws profile that contained all important information like secret key, region, path to bucket etc. Anyways I established a connection to the database and locally I was able to withdraw information from the database. My app was running fine locally. However the app when uploaded to shiny app io and deployed a second time it would not show updated dates in table. I thought this might be due to how I set up the scripts in my app. It now contains three scripts only ui.R, server.R and a global.R However after I did this I kept getting the error:

Error: Boto3 is not detected please install boto3 using either: pip install boto3 in terminal or install_boto(). Alternatively reticulate::use_python or reticulate::use_condaenv will have to be used if boto3 is in another environment.

So I did as the error said to and I added:reticulate::use_virtualenv("RAthena") to my global.R script where all data withdrawals and calculations are done. Here is the stack overflow answer. After I did this I got the error:

Error: Unable to find conda binary. Is Anaconda installed?

I decided to try then reticulate::use_condaenv("RAthena") I then simply got the boto3 error again. The annoying part is I have checked in my terminal and Boto3 is successfully installed. I also have Anaconda installed I even updated Boto3.

I saw from the answer that it is possible to use noctau to establish a database connection. I did this and everything worked fine locally. I was able to withdraw data from the database. Of course when I deployed this to shiny app io I got the error

Error in value[3L] : No region provided

I don't know why I have this error. In my aws profile my region for the database is specified. In R studio I see it in the connections tab in the right hand corner. With the code dbGetInfo(con) I see it there as well. I don't know what the solution is.

Answer 1

For RAthena it appears shinyapp.io is having difficulty setting up your environment. This tutorial may help solve the RAthena problem. Tutorial: using Shiny + reticulate to create apps with R and Python 3

For noctua, it looks like it is having difficulty getting your aws credentials when on shinyapp.io server. When setting up the shinyapp.io how have you set up your credentials for AWS? Are you using environmental variables or are you using .aws/credentails and .aws/config files?

You might want to consider the config package recommended in the shinyapps.io user guide book: https://docs.rstudio.com/shinyapps.io/applications.html#config-package

Here is an example of how to set it up.

Setting up config.yml

local:
  dataconnection:
    profile: 'my_personal_profile'
    aws_access_key_id: 'my_aws_key'
    aws_secret_access_key: 'my_aws_secret_key'
    region: 'eu-west-1'
    database: 'default'
    s3_bucket: 's3://my/s3/bucket/'
    
shinyapps:
  dataconnection:
    profile: 'shiny_profile'
    aws_access_key_id: 'shiny_aws_key'
    aws_secret_access_key: 'shiny_aws_secret_key'
    region: 'eu-west-1'
    schema: 'default'
    s3_bucket: 's3://shinyapp/bucket/'

library(DBI)
conn_args <- config::get("dataconnection")
con <- dbConnect(noctua::athena(),
   profile_name = conn_args$profile,
   aws_access_key_id = conn_args$aws_access_key_id,
   aws_secret_access_key    = conn_args$aws_secret_access_key,
   region_name    = conn_args$region,
   schema_name   = conn_args$schema,
   s3_staging_dir = conn_args$s3_bucket
)

When the environment variable R_CONFIG_ACTIVE=local, then the local credentials will be used. When R_CONFIG_ACTIVE=shinyapps (on shinyapps.io), the production credentials will be used.