In robertzk/berdie: Berdie

library(berdie)

Loading a database connection

Database queries produce results in table format -- a certain number of rows by columns. This is essentially a data.frame, so it makes sense to be able to turn queries into data.frames given a database connection.

A database connection, like in Rails, is given by a database.yml file. You can pass in the file explicitly to postgresql_connection:

connection <- postgresql_connection("some/dir/database.yml")

However, it is easier to let the package do its job and set either

• The berdie.database.yml option with the location of the configuration file. You can ensure this always runs on startup by placing it in your ~/.Rprofile file with options(database.yml = "some/dir/database.yml").

• Load the syberia package. If a syberia project is already loaded, the current value of syberia_root() will be used and the relative file config/database.yml will give the database configuration. Thus, if your syberia project is at some/dir, the default connection used by Berdie will be the one given by some/dir/config/database.yml.

Running a query on a database

Once you have an open connection, running queries is straightforward:

df <- run_query("select * from some_table limit 5", connection)
# A data.frame with (at most) five rows.

However, you do not have to lug the connection around.

df <- run_query("select * from some_table limit 5")

The default is to use the helper function last_connection(), which will attempt to look for: (1) the previous connection used for a query execution, (2) load the connection from the berdie.database.yml global option, or (3) borrow the database.yml file from a syberia project.

Whatever the case, you can always inspect the connection used by the previous query by asking for last_connection(). If you place library(syberia) and library(berdie) in your ~/.Rprofile and you have set a default syberia project, you should be able to execute run_query on starting R without any setup.

robertzk/berdie documentation built on May 27, 2019, 10:34 a.m.