FhirJsonParser Removing Leading Spaces

[MLuebbers]

I have a FHIR JSON string I wish to deserialize to a Resource object. There is an identifier in the resource that has a leading space and must remain that way but when inspecting the resulting Resource POCO from the parser that space is gone. I was under the impression the parser should leave JSON string fields exactly as they are without removing any leading or trailing whitespace. The SerializationSettings by default seem to remove the whitespace but I couldn't find an equivalent for the ParserSettings. I couldn't find any documentation about this behaviour, is this intended?

[alexzautke]

Hi @MLuebbers, we apply the same rules as for XML here: "Implementers SHOULD trim leading and trailing whitespace before writing and SHOULD trim leading and trailing whitespace when reading attribute values (for XML schema conformance)" (see https://build.fhir.org/xml.html). There is no setting for this as it confuses all kinds of tooling.

[alexzautke]

Sorry, there are currently only limited development resources for this. We'll try to take a look at this in the next weeks. In the meantime you could add an extension like https://build.fhir.org/ig/HL7/fhir-extensions/StructureDefinition-originalText.html on the element to capture its original intent.

[mikemassa84]

I found that updating FhirJsonBuilder.buildValue and removing the trim statement fixes our issue. Not sure if you want/need to add a setting for that value.

        private JValue buildValue(object value, string requiredType = null) => value switch
        {
            bool or decimal or Int32 or Int16 or ulong or double or BigInteger or float => new JValue(value),
            string s => new JValue(s), // string s=> new JValue(s.Trim()),
            long l when requiredType is "integer" or "unsignedInt" or "positiveInt" => new JValue(l),
            _ => new JValue(PrimitiveTypeConverter.ConvertTo<string>(value)),
        };

I found this comment in the BaseFhirJsonPocoSerializer regarding not trimming string values so maybe we don't need a serializer setting for that?

                // A little note about trimming and whitespaces. The spec says:
                // "(...) In JSON and Turtle whitespace in string values is always significant. Primitive types other than
                // string SHALL NOT have leading or trailing whitespace."
                // Based on this, we are not trimming whitespace here. Validation is not a part of the responsibilities of
                // the serializer, and string-based types (like code and uri) should make sure their values are valid,
                // so should not have trailing spaces to begin with. strings are allowed to have trailing spaces, but should
                // not just be spaces. The serializer will, however, not serialize an element with only whitespace
                // (or an empty byte[]).
                case string s: writer.WriteStringValue(s); break;

[mikemassa84]

I created PR #3273 with a new serializer setting if that helps.

[andrzejskowronski]

Thank you for the contribution, that indeed does make it a bit quicker 🙂

[mikemassa84]

The latest snapshot with both changes is working for us. Thanks for all the help and the quick turnaround!