Open a dataset in arrow format

Since the replay datasets are so large (often > 1 million rows x thousands of columns), they cannot be read into R directly. It works better to keep the data out of memory and only load the necessary parts. This function provides the capability to do this using the arrow package.

Usage

mr_open_17lands_arrow(
  set,
  data_type,
  event_type,
  use_cache = TRUE,
  overwrite = FALSE,
  cache_dir = mr_cache_dir(),
  quiet = FALSE
)

Arguments

set

Three-letter set code. Not case-sensitive. For a complete list of set names and codes, run mr_get_sets() (not all sets are available on 17lands).

data_type

Type of data. Must be one of "game", "draft", or "replay".

event_type

Type of event. Must be one of "premier", "traditional", "sealed", or "quick". Not all formats are available for all event types (in particular, Quick Draft is only available for one set at the time of writing).

use_cache

Logical; should the cache be used?

overwrite

Logical: should any existing file be overwritten?

cache_dir

Path to the folder used for caching data.

quiet

Logical: should messages emitted by this function be silenced?

Value

Character vector

Details

Not all sets are available. To see the available files, visit https://www.17lands.com/public_datasets.

Currently only works for 'replay' data.

Examples

# See what happened on user's turn 1 for all KTK games
library(dplyr)
#> 
#> Attaching package: ‘dplyr’
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, lag
#> The following objects are masked from ‘package:base’:
#> 
#>     intersect, setdiff, setequal, union
ktk_replay <- mr_open_17lands_arrow("KTK", "replay", "premier")
ktk_turn_1 <- ktk_replay |>
  select(
    draft_id, game_number, contains("user_turn_1_")) |>
  collect()