If thats the case then, in a Mapper Snap(befor the JSON Splitter), parse the String to JSON using JSON.parse function.
@Raph, I assume you want to do some transformation to the data, or otherwise, I don’t think you need the mapper at all; just push the data as-is to the target, without the mapper.
If it is the case that you need the mapper and want to maintain the key name order in the output data (and don’t necessarily care about the schema view), you may get by with just checking the checkbox in your mapper. It won’t re-order the schema view in the UI, but you can still manipulate some or all of the data and the output to the target should return to original order.
Otherwise, if you absolutely need the schema view in the UI to represent the original order, attached below is a sample pipeline with an optional hack to help.
The Pad Key Names mapper prefixes the key names with spaces, keeping them all the same fixed length, based on the length of the rightmost column key name. Since spaces are first in the schema sort order, and all your key names should be upper case, the Transform Values mapper schema should keep the same key order in its input schema, while allowing you to manipulate the data for the target. The Trim Key Names mapper reverts the key names back to their original names, without spaces, while maintaining the key order.
Community.8623_2020_10_25.slp (8.1 KB)
