I think you need a short section on managing credentials in between "run queries safely" and "deploying content". This would basically say that you should never store a password in your R script and never type it into the console.

For now, this section would discuss three options to avoid this:

• Use rstudioapi::askForPassword()
• Use keyring
• Use an environment variable

Under the Availability header the code block for installing dbplyr reads:

devtools::install_github("tidyvers/dbplyr")

It should instead read:

devtools::install_github("tidyverse/dbplyr")

@hannesmuehleisen - I think that we should include MonetDB and MonetDBLite-R to the Databases section. Any suggestions or objections to that? If none, would you have any recommendations for connecting that are different from what in the README in the repo (https://github.com/hannesmuehleisen/MonetDBLite-R)?

The following links, which can be found on https://db.rstudio.com/best-practices/dashboards/#full-example, are broken:

• local_app.R
• db_app.R

Look into updating the ODBC drivers installation instructions for Linux: rstudio/rstudio#2463

Update http://db.rstudio.com/databases/big-query/ with info from the updated README in bigrquery: r-dbi/bigrquery#202

Expand on the ideas on this thread:

https://community.rstudio.com/t/why-does-rstudio-documentation-recommend-odbc-vs-jdbc-drivers/2381/5

cc @hadley

@carneybill

Add:

{{ template "_internal/google_analytics.html" . }}

To a new footer-custom.html file inside the layouts/partial folder

This is something that comes up whenever I talk about databases, and it would be good to have "official" guidelines that we can direct people to. Since DBI and dplyr overlap a lot in their database functionality, users (especially newbies) are often at a loss on the topic of which one to choose.

In the olden times, when dplyr only supported SELECT-power statements, the rule was easy. Now, not so much. However, that may still be a good rule of thumb as non-SELECT statements in dplyr have not been time-tested yet.

Whatever the official recommendation is, we should have one, so that the answer to "Which package to use?" is not completely the opinion of whomever happens to be answering it.

cc @hadley

https://github.com/rstudio/dbtest#yaml-file

Put together an article that discusses steps to take during and after a Shiny app that uses a database as it's source is built. Using an example, the plan is to show how to use the following tools to improve the app's performance:

• The pool package
• Use profvis, and possibly shinyloadtest, after the app is complete
• How to use show_query() and explain() to troubleshoot long running queries

The article should also include links to these pages:

• https://support.rstudio.com/hc/en-us/articles/231874748-Scaling-and-Performance-Tuning-in-RStudio-Connect
• http://shiny.rstudio.com/articles/#performance
• http://shiny.rstudio.com/articles/#profiling

@bborgesr
@jcheng5

https://twitter.com/joelgombin/status/928195603947048960

It looks like we need to add connection pages, such as this one for MS SQL (https://db.rstudio.com/databases/microsoft-sql-server/), for the new drivers we offer:

1. Apache Cassandra
2. Amazon Athena
3. MongoDB
4. Google BigQuery
5. IBM Netezza

Can you add examples as comments to this Issue? I'll be glad to create each page

Add a section to Impala about cleaning up your metadata. If you create a table in Impala and then drop the Hive metadata, you will need to invalidate the Impala metadata.

impala_con <- dbConnect(odbc::odbc(), "Impala")
dbWriteTable(impala_con, "mtcars", mtcars)
hive_con <- dbConnect(odbc::odbc(), "Hive")
dbRemoveTable(hive_con, "mtcars")
dbReadTable(impala_con, "mtcars") # succeeds
dbExistsTable(impala_con, "mtcars") # fails
dbGetQuery(odbcCon, "INVALIDATE METADATA mtcars")
dbExistsTable(impala_con, "mtcars") # succeeds

This happens because dropping the Hive metadata does not drop the Impala metadata. More information can be found here.

^^ Should have been self evident :p

https://www.rstudio.com/products/drivers/

I think it would be really helpful, if there was an article about adding and more importantly updating existing records in a database through R, and the best practices to use. There are many excellent resources on downloading, cleaning and transforming data in R, but data warehousing is a crucial piece missing in the data pipeline.

Add a table that lists supported data types for each data source. For example:

Supported Teradata Data Types

Unsupported types will throw an error. See r-dbi/odbc#238.

MySQL does not have a boolean type and uses TINYINT instead. Therefore, if you write logical values they will get returned as integers. Please make a note in the documentation.

Provide some FAQ level answers to common issues users have reported when connecting and using a database. It should consist mostly of links to other articles, such as:

• https://support.rstudio.com/hc/en-us/articles/115011860827-System-Requirements-for-ODBC-Connections-and-RStudio-Professional-Drivers

And links to the Known Issues sections of those databases we have documented issues for.

The main idea is to have a very visible "Troubleshooting" link in the page that users can go in. We can update it as users report to use that they couldn't find what they were looking for.

Another idea is to include the link to community.rstudio.com to also cover non-customers who are still having issues.

@jimhester
@nwstephens

Add the Environment Variables feature in RSC as an option in this page: http://db.rstudio.com/best-practices/portable-code/

Ref: https://blog.rstudio.com/2018/04/12/rstudio-connect-1-6-0-a-year-in-the-making/

And instead use named parameters to dbConnect()

Using https instead of http makes the site render incorrectly

https://db.rstudio.com/

http://db.rstudio.com/

Right now we have "Best Practices" as a section header, but no other section headers, so the sidebar feels inconsistent. We should change the structure to something like this:

I added "Examples" and "Packages" as section headers

This should be a relatively simple change to config.toml

The content of this page: http://db.rstudio.com/advanced/snippets/ , needs to be sourced from here: https://rstudio.github.io/rstudio-extensions/rstudio-connections.html

The current pool page doesn't go into much detail about where and when to use different pool functions.

For example:

• where/when in your shiny app to create a pool object
• where/when to checkout connections from the pool with poolCheckout()
• where/when to when to return connections to the pool with poolReturn()
• where/when to close the pool using poolClose()

The current page also only has guidance when using dplyr to query, but not using custom SQL.

Addresses: https://community.rstudio.com/t/pool-and-shiny-best-practices-for-getting-custom-queries-from-database/3707/4

Ref tidyverse/dplyr#3343

Add page to explain how to switch between databases using dbplyr and DBI/odbc.

It should include how to use our functions and function arguments, such as in_schema() and dbListTables(...schema_name=""), and some DB specific commands like Oracle's SET SCHEMA. Also, cross reference those DB specific commands to their respective page in the site.

cc.
@jimhester
@nwstephens
@hadley

cc @nwstephens

Looks like there's a broken link on this page - https://db.rstudio.com/rstudio/connections/

• the link to "Integrated Security with DSN" goes to - http://127.0.0.1:4321/managing-credentials/#integrated-security-with-dsn
  This should instead go to https://db.rstudio.com/best-practices/managing-credentials/#integrated-security-with-dsn

In the /dplyr page

cc @krlmlr

https://community.rstudio.com/t/dbi-dbwritetable-is-slow-for-writing-to-azure/19262/21

Hi

I was looking to correct an error but it already has been corrected here cddf2dc

but it is not yet online.
https://db.rstudio.com/databases/postgresql/

html file has not been rebuilt and deployed

cheers

See: r-dbi/RPostgres#201

• Add reference to dbplot in the best practice visualization article
• Add dbplot to the packages page

r-dbi/bigrquery#202

Under the "Connection Settings" header, then the "Driver" bullet.

In "See the Drivers section for setup information", the word Driver is a hyperlink to https://db.rstudio.com/redshift/drivers which loads a "Page Not found" page.

It is not entirely clear to me if the expected behavior is to lead to the anchor tag above https://db.rstudio.com/redshift/#driver-options or if there was supposed to be a page in the location it links to.

The PostgreSQL has identical behavior, except the link that isn't found is https://db.rstudio.com/postgresql/drivers and I do not seem to see an anchor tag on the page that would make sense for it to jump to.

Most references are pointed to RPostgreSQL package, need to make sure all are pointed to that package instead of RPostgres