I have json input of form as shown below. It can also be empty like []
[
{
"NAME": "Aron",
"CITY": "NewYork",
"PROV": "NY",
"POSTALCODE": "12345",
"DATE": "2021-08-19",
"TIME": "14:25:25"
},
{
"NAME": "Paul",
"CITY": "Chicago",
"PROV": "BI",
"POSTALCODE": null,
"DATE": "2021-08-19",
"TIME": ""
}
]
I am writing NiFi JoltTransformJSON spec so my output will contain the JSON as
[
{
"NAME" : "Aron",
"Address" : "NewYork, NY, Pin: 12345",
"DATE" : "2021-08-19",
"TIME" : "14:25:25"
},
{
"NAME" : "Paul",
"Address" : "Chicago, BI",
"DATE" : "2021-08-19",
"TIME" :""
}
]
Basically it merges City, Prov, Postalcode and puts in new key Address. They are concatenated using , comma between them and also if Postalcode is not null it is concatenated in Address as Pin:. If everything (City, Prov, Postalcode) is null then Address will be empty like "Address":""
I am trying with shift followed by remove
[
{
"operation": "shift",
"spec": {
"@(1,CITY)": "",
"*": "&"
}
},
{
"operation": "remove",
"spec": {
"CITY": "",
"PROV": "",
"POSTALCODE" : "",
}
}
]
I have achieved this using SplitJSON, EvaluateJSONPath and UpdateAttribute. But I am trying this using JoltTransformJSON now. Is this achievable using JoltTransformJSON ?
You can consecutively use modify-overwrite-beta along with concat and size functions and shift transformations such that
[
{
"operation": "modify-overwrite-beta",
"spec": {
"*":{
"Address1": "=concat(@(1,CITY),',',@(1,PROV))",
"Address2": "=concat(',Pin:',@(1,POSTALCODE))",
"Addr2Size": "=size(@(1,Address2))",
"Address": "=concat(@(1,Address1),@(1,Address2))"
}
}
},
{
"operation": "shift",
"spec": {
"*":{
"NAME": "[&1].&",
"Addr2Size": {
"5": {
"@(2,Address1)": "[&3].Address"
},
"*": {
"@(2,Address)": "[&3].Address"
}
},
"DATE": "[&1].&",
"TIME": "[&1].&"
}
}
}
]
where size function is used to determine whether POSTALCODE value is null or not.