ParseException on valid JSONata with parameter as key in object constructor

[robertvanloenhout]

Given any input and the following JSONata:

(
$myfunction := function ($key, $value) {
  {
    $key: $value
  }
};
$myfunction('key', 'value')
)

This produces the following error:

Exception in thread "main" com.api.jsonata4java.expressions.ParseException: line 4:4 at [@11,51:54='$key',<41>,4:4]: no viable alternative at input 'function($key,$value){{$key'

It produces no error on https://try.jsonata.org/
It seems that a param as key for object constructors is not allowed, but it should be.

[aubelsb2]

Just ran into exactly the same issue

[wnm3]

Will try to fix tomorrow.

[wnm3]

This error is related to the parser not recognizing a variable as the key in an object. I'll try to get this fixed but parser errors have ripple effects so it may take some time. Referencing the variables as values is not an issue:

Using:
( $myfunction := function ($key, $value) { {'a':$key,'b':$value} }; $myfunction('key', 'value') )
Results in:
{
"a" : "key",
"b" : "value"
}

[aubelsb2]

Thanks!

[wnm3]

This was a good bug to fix. The parser was only expecting a String as a key so I changed it to allow for an expression that resolves to a string. I also added some runtime exceptions to match jsonata.org behavior. There is an odd behavior where an errant value just causes the key:value pair to be dropped from the object altogether -- not sure if this is a bug or feature of jsonata.org but I did the same thing.

This caught a stupid bug of mine as well when constructing objects where I missed the comma field delimiter for the first pair of fields.

I've added your test case to the BasicExpressionsTest as well so we'll continue to check that we don't break it in the future.

There was a parse exception being tested for:

{`hello`:"world"}

But now this will parse. However, it throws the same runtime exception as jsonata.org:
Key in object structure must evaluate to a string; got: undefined

Anyway, I've released version 1.5.1 at 5:20pm ET so in an hour or so, Maven Central should have the updated jar file, and I checked in these changes against this issue if you are interested to see what was involved.

Please close this issue if you believe it is fixed.

[robertvanloenhout]

I've updated the JSONata4Java dependency to 1.5.1, and the issue has been resolved.
Thanks for fixing it!