In seananderson/ramlegacy: Download and Cache the RAM Legacy Stock Assessment Database

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Import the RAM Legacy Stock Assessment Database into R

Please use Carl Boettiger and Gupta Kshitiz's ramlegacy package https://github.com/ropensci/ramlegacy instead.

This package does one thing: it downloads a Microsoft Access copy of the RAM Legacy Stock Assessment Database and converts it to a local sqlite3 database. This makes it easy to work with dplyr/dbplyr, for example. The make_ramlegacy() function also leaves a copy of .csv files for each table in the database in the R working directory if you'd prefer to work with those.

Note that you must have the utility mdb-tables installed and in your path from mdbtools. This utility provides tools for extracting Access databases. You can find installation instructions at http://mdbtools.sourceforge.net. If you're on OS X and using homebrew, you can install it with brew install mdbtools.

Example use

library("knitr")
opts_chunk$set(cache=TRUE)

Install the package:

# install.packages("devtools")
devtools::install_github("seananderson/ramlegacy")

Cache and convert the database:

library("ramlegacy")
make_ramlegacy()

Work with the data:

library("dplyr")
ram <- src_sqlite("ramlegacy.sqlite3")
ram # see the available tables

Access the area table:

tbl(ram, "area")

Join the time series ts and stock tables on the stockid column:

ts <- tbl(ram, "timeseries")
stock <- tbl(ram, "stock")
select(stock, stockid, scientificname, commonname, region) %>%
  inner_join(ts)

Note that because you are working with dplyr and a database, you will need to use dplyr::collect() once you want to load the data into a local data frame. For example:

tbl(ram, "area") %>% 
  collect()

For safety, you may want to use dplyr::collect(n = Inf) to return all rows of data, not just the minimum default number. In this case it won't make a difference.

tbl(ram, "area") %>% 
  collect(n = Inf)

seananderson/ramlegacy documentation built on May 12, 2020, 11:41 a.m.