read_gesla {geslaR}

Read a GESLA dataset

Description

Read a CSV or Parquet file, as exported from the GESLA Shiny app interface (geslaR-app). A "GESLA dataset file" is a subset of the GESLA dataset, fetched from the geslaR-app. When using that app, you can choose to download the selected subset in CSV or Parquet file formats. Whichever option is chosen this function will automatically identify the file type and use the appropriate functions to import the dataset to R.

This function can be used for exported files from the online interface (hosted in this server) or from a local interface, as when using the run_gesla_app() function.

Usage

read_gesla(file, as_data_frame = FALSE, ...)

Arguments

file

The file name (must end in .csv or .parquet only)

as_data_frame

If FALSE (default), the data will be imported as an Arrow Table format. Otherwise, the data will be in a tbl_df (data.frame) format. See Details.

...

Other arguments from arrow::read_csv_arrow(), and arrow::read_parquet(), from the arrow package.

Details

We highly recommend to export subsets of the GESLA dataset from the geslaR-app in the Parquet file format. This format has a much smaller file size when comparred to the CSV format.

In any case, the only difference between CSV and Parquet files will be the file size. However, when importing these data to R, both file types have the option to be imported as an Arrow Table format, which is the default (argument as_data_frame = FALSE). This way, the object created in R will have a very small size, independent of how big the file size is. To deal with this type of object, you can use dplyr verbs, in the same way as a normal data.frame (or tbl_df). Some examples can be found in the Arrow documentation.

If the as_data_frame argument is set to TRUE, the imported R object will vary in size, according to the size of the dataset, and regardless of the file type. In many situations, this can be infeasible, since the object can result in a "larger-than-memory" size, and possibly will make R operations slow or even a session crash. Therefore, we always recommend to start with as_data_frame = FALSE, and work with the dataset from there.

See Examples below.

Value

An Arrow Table object, or a tbl_df (data.frame)

Author(s)

Fernando Mayer [e-mail]

Examples

##------------------------------------------------------------------
## Import an internal example Parquet file
tmp <- tempdir()
file.copy(system.file(
    "extdata", "ireland.parquet", package = "geslaR"), tmp)
da <- read_gesla(paste0(tmp, "/ireland.parquet"))
## Check size in memory
object.size(da)

##------------------------------------------------------------------
## Import an internal example CSV file
tmp <- tempdir()
file.copy(system.file(
    "extdata", "ireland.csv", package = "geslaR"), tmp)
da <- read_gesla(paste0(tmp, "/ireland.csv"))
## Check size in memory
object.size(da)

##------------------------------------------------------------------
## Import an internal example Parquet file as data.frame
tmp <- tempdir()
file.copy(system.file(
    "extdata", "ireland.parquet", package = "geslaR"), tmp)
da <- read_gesla(paste0(tmp, "/ireland.parquet"),
    as_data_frame = TRUE)
## Check size in memory
object.size(da)

##------------------------------------------------------------------
## Import an internal example CSV file as data.frame
tmp <- tempdir()
file.copy(system.file(
    "extdata", "ireland.csv", package = "geslaR"), tmp)
da <- read_gesla(paste0(tmp, "/ireland.csv"),
    as_data_frame = TRUE)
## Check size in memory
object.size(da)

## Remove files from temporary directory
unlink(paste0(tmp, "/ireland.parquet"))
unlink(paste0(tmp, "/ireland.csv"))