Database functionality

[sckott]

Database functionality

[seankross]

I think we need to include functionality for setting up a database. We should make as many sane choices for users as possible, and we should assume that users are going to use individual databases for basic data science tasks. Therefore any hyper-scale database functionality is out of scope

[sckott]

Okay, can include db setup function(s).

[sckott]

Working on this on branch dbs

notes trying to sort out how to do these functions

• https://aws.amazon.com/blogs/opensource/getting-started-with-r-on-amazon-web-services/
• https://blog.davisvaughan.com/posts/2017-05-10-aws-rds-r/
• https://medium.com/analytics-vidhya/connecting-amazon-relational-database-with-r-d9b800e5efb0
• connecting with a python client https://docs.aws.amazon.com/redshift/latest/mgmt/python-connect-examples.html
• sorting out connection failure https://docs.aws.amazon.com/redshift/latest/mgmt/connecting-refusal-failure-issues.html
• localstack docs for redshift https://docs.localstack.cloud/user-guide/aws/redshift/

[sckott]

Right now I made the following assumptions/decisions/questions:

• no fxns for deleting/updating/etc. databases are available in this pkg. I assume we just provide the creation fxn, and then the user is left to use the AWS dashboard to do the rest. Do we want to provide additional fxn-ality beyond create and connect?
• there's only one client for each of RDS and Redshift right now. That is, one set of credentials for each of the services. The only issue comes if a user has 2 or more sets of credentials they want to use in the same R sessions. See #20 for more
• we do not have fxns for setting up security groups right now. should we? then we could make it one less step outside of R to get a DB up and running?
• generate strings for users? right now to set up a DB users have to think of an ID name for the instance, and set a user name and password. We could use a dependency to generate these things. Should we?

@seankross