Need to change csv header from camel case to snake case in dataweave

I am working on a pipeline to dynamically dump all columns from the salesforce object to the S3 bucket. I get all columns for a salesforce object using describe object API. I store all columns into a variable and then create a big SOQL query out of it and submit a bulk query job v2.

Now, this is the main problem. The Column name I am getting from the salesforce connector is in camelCase

 [{
    "Id": 123, 
     "FirstName": "Manual", 
     "MasterRecordId__c" :"abc"
    },
    {   
    "Id": 456, 
     "FirstName": "John", 
     "MasterRecordId__c" :"def"
 }] 

But I want column names to be in snake case

[{ 
   "Id": 123, 
   "first_name": "Manual", 
   "master_record_id__c":"abc"
  },
  {   
   "Id": 456, 
   "first_name": "john", 
   "master_record_id__c":"def"
 }] 

I understand mulesoft has an underscore function to do the same thing, but I am not able to apply any function at "key" level.

Any lead would be really helpful. Please let me know for any questions.

You just have to use mapObject along with the underscore function

%dw 2.0
import underscore from dw::core::Strings
output application/json
---
payload map ((item) -> 
    item mapObject ((value, key) -> {
        (underscore(key)): value       
    })
)