Mapping JSON Data Types
Integration Server maps JSON data types to IS data types as follows:
JSON data type | IS data type |
object | Document |
string | String |
number (real) | Float or Double Java wrapper, depending on the option you select for mapping real numbers. |
number (integer) | Integer or Long Java wrapper, depending on the option you select for mapping integers. |
true/false | Boolean Java wrapper type |
array | Array of an Integration Server data type.  If the JSON object contains a string array, Integration Server sets the corresponding field in the IS document type to String List. If the JSON object contains a two-dimensional string array, Integration Server sets the corresponding field in the IS document type to String Table. If the JSON object contains a array having different types of objects (example, integer, string), Integration Server sets the corresponding field in the IS document type to Object List with Java wrapper type as UNKNOWN. Note:
The Flow type system in Integration Server does not support the following multi-dimensional arrays in JSON objects: Two-dimensional arrays of data types other than string Multi-dimensional arrays of three or more dimensions (three-dimensional, four-dimensional, and so on.) For IS document types created from a JSON object using Designer, these kinds of arrays become an Object with a Java wrapper type of UNKNOWN. The JSON coder in Integration Server can create pipelines with arbitrarily deep arrays or various types and you can write Java services to access those arrays. The limitations above apply to Flow services and viewing the document in Designer. |
array of null | Object List |
null | null |
All others | String Note:
If an object has a toString() implementation, Integration Server uses that implementation. If the object does not provide a toString() implementation, Integration Server uses Object.toString(). Object.toString() returns the class name and hexadecimal representation of the hash code of the object, such as "javax.namining.InitialContext@3ae6f00b". |
Additional Notes on Data Conversion
If JSON text begins with an array at the root and the array is unnamed, when parsing the JSON text,
Integration Server uses a fixed name of
$rootArray for the array value. The
$rootArray field appears in the output pipeline of the service converting the JSON text to a Document (IData). When creating a JSON response, if the pipeline contains
$rootArray with an array as its value,
Integration Server discards the
$rootArray name and transforms the array value into a JSON array. The resulting JSON text will contain only an unnamed array.
If JSON text contains only a primitive type (string, number, boolean, or null) at root level,
Integration Server uses a fixed name of
$rootValue for the primitive value while converting it to IData.
The following table identifies how JSON text with a primitive type is converted to an IData in Integration Server:
JSON text | Generated IData |
"just a simple string" | {
$rootValue="just a simple string"
} |
3.14159 | {
$rootValue=3.14159
} |
true | {$rootValue=true} |
null | {$rootValue=null} |
null | {} Empty IData (if decodeNullRootAsEmpty input parameter is set to true) |
With regards to using
Integration Server data types with JSON,
Integration Server supports only those types that can be mapped to a JSON value as defined in
https://tools.ietf.org/html/rfc7159#section-3.
Integration Servercan take any valid arbitrary JSON text and convert it to an IData.
Integration Server must be able to convert the resulting IData to JSON text that is identical to the original text. If
Integration Server cannot do that for an
Integration Server data type, then
Integration Server does not support the use of that data type with JSON. For example, com.wm.util.Table is not supported for JSON even though it is supported for XML.
Integration Server embeds additional type information in XML when converting IData to XML. However,
Integration Server cannot embed the additional type information in JSON because the additional type information is treated as JSON text. The resulting JSON text would not match the original JSON text.
An exception to the guideline that
Integration Server must be able to convert the resulting IData to JSON text that is identical to the original text involves primitive JSON type null. If the JSON text is null and
decodeNullRootAsEmpty is set to true in the service performing the conversion, JSON text of null will get converted to empty IData and an empty IData will get converted to {} JSON text.