Support new field type list

[peterdesmet]

CHANGELOG: https://datapackage.org/overview/changelog/#list-field-type-new

• Decide on support, e.g. don't support this (i.e. default to string) or convert values on delimiter to a vector.
• Update documentation
• Update tests
• See also suggestions in Support delimited arrays #173

[khusmann]

I'm interested in this field type for representing multiselect items (although it will have to wait until the list field type can be extended by the categories property)

In the meantime, I'd vote for the latter approach (convert the cell to a vector using delimiter and load as list-columns).

For example, the csv:

row_id, field1
1, "a,b,c"
2, "d,e"
3, "f"

with schema fields:

[
  {
    "name": "row_id",
    "type": "integer"
  },
  {
    "name": "field1",
    "type": "list",
    "delimiter": ",",
    "itemType": "string"
  }
]

would become:

library(tidyverse)
tibble(
  row_id = 1:3,
  field1 = list(
    c("a", "b", "c"),
    c("d", "e"),
    c("f")
  )
)
#> # A tibble: 3 × 2
#>   row_id field1   
#>    <int> <list>   
#> 1      1 <chr [3]>
#> 2      2 <chr [2]>
#> 3      3 <chr [1]>

^{Created on 2024-06-07 with reprex v2.0.2}