F# - JSON Type Provider - Literal String - f#

Question

I have an interesting problem using the F# Type Provider. Using the following sample json to generate the required type works beautifully but with one glaring issue - one of the values is a literal string ("235") but that string does not occur in the real data - each new record has a different three-digit number:
{
"record": {
"235": {
"Id": "001",
"Name": "A. N. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00"
...
}
"255": {
"Id": "005",
"Name": "D. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00"
...
}
}
So, unfortunately, I cannot access the Id, Name, and IdDateTime fields without knowing the three-digit code in advance!
Any advice?

Answer 1

Do you have to use the JsonProvider? There is a simpler JsonParser in FSharp.Data that allows easier access to more dynamic data. After that it's just a simple exercise to destructure it to get to the inner part.
#r #"../packages/FSharp.Data/lib/net40/FSharp.Data.dll"
open FSharp.Data
open FSharp.Data.JsonExtensions
[<Literal>]
let jsonText = """
{
"record": {
"235": {
"Id": "001",
"Name": "A. N. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00"
}
}
}
"""
let json1 = JsonValue.Parse(jsonText)
let json2 = json1?record
let json3 =
match json2 with
| JsonValue.Record x -> x
let json4 =
match json3 with
[|(_,x)|] -> x
val json4 : JsonValue = { "Id": "001", "Name": "A. N. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00" }
Edit
On an array of jsonvalues this works the same, except you have to map the matching function over. For example:
let jsonText2 = """
{
"record": {
"235": {
"Id": "001",
"Name": "A. N. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00"
},
"255": {
"Id": "005",
"Name": "D. Other",
"IdDatetime": "2017-11-11T13:10:00+00:00"
}
}
}
"""
let json1 = JsonValue.Parse(jsonText2)
let json2 = json1?record
let json3 =
match json2 with
| JsonValue.Record x -> x
let json4 =
json3
|> Array.map (function _,x -> x)
val json4 : JsonValue [] = [|{ "Id": "001", "Name": "A. N.
Other", "IdDatetime": "2017-11-11T13:10:00+00:00" };
{ "Id": "005", "Name": "D. Other", "IdDatetime": "2017-11-11T13:10:00+00:00" }|]