Skip to contents

Parse and validate JSON in one step

Source: R/serialize.R
from_json.Rd

JSON arrays come back as R vectors and arrays of objects as data frames, because the declared types accept those shapes – there is no separate JSON mode. Reshaping happens only here and in from_list(), never when a value is handed straight to a type.

Usage

from_json(t, txt)

Arguments

t

A type, or anything as_type() accepts.

txt

A JSON string, or anything jsonlite::fromJSON() reads.

Value

The validated value.

See also

to_json(), from_list(), try_parse()

Examples

Reading <- struct("Reading", station = chr[1], celsius = num[1], at = datetime[1])
from_json(Reading, '{"station": "KOA", "celsius": 21.5, "at": "2024-05-17T09:30:00"}')
#> <Reading>
#>   station : "KOA"
#>   celsius : 21.5
#>   at      : 2024-05-17 09:30:00

from_json(map_of(int[1]), '{"cpu": 4, "memory": 16}')
#> $cpu
#> [1] 4
#> 
#> $memory
#> [1] 16
#> 
from_json(frame(id = int, nm = chr), '[{"id": 1, "nm": "a"}]')
#>   id nm
#> 1  1  a

try(from_json(Reading, '{"station": ["a", "b"], "celsius": 21.5, "at": "nope"}'))
#> Error : 2 validation problems in Reading
#>   $station  expected chr[1], got list[2]  -- length is 2, not 1
#>   $at       expected datetime[1], got chr[1] "nope"  -- expected YYYY-MM-DDThh:mm:ss