Extracting Json from a Json Data in Hive - json

Question

I have the following data which is in the json format in a column called details of table named customer in Hive:
{
"customer" : {
"given_name" : "Anuvrat",
"surname" : "Singh"
},
"order" : {
"id" : "123dfe523gd"
},
"address" : {
"city" : "kolkata",
"pin" : "700091"
},
"phone" : {
"mobile" : "*********"
}
}
I have to remove the address and phone from the json data and the data should look like:
{
"customer" : {
"given_name" : "Anuvrat",
"surname" : "Singh"
},
"order" : {
"id" : "123dfe523gd"
}
}
How to do(i.e update) for every row present in the table?
I tried the following command hadoop fs -cat /home/customer/* | jq '.details[] |= del(.address,.phone)' yet I dint get the expected output rather getting error saying
parse error: Invalid numeric literal at line 1, column 93
cat: Unable to write to output stream.

Answer 1

If you're open to a solution not using Hive, I want to remark that this is something very easy to do with jq command line JSON parser.
Given your input file, you would do:
jq 'del(.address,.phone)' file
If you want to remove address and phone objects for all entries of the table, you can do:
jq '.[] |= del(.address,.phone)' file

Answer 2

This is the query which I ran to get the above result:
INSERT OVERWRITE TABLE customer Select id,CASE WHEN id is not null THEN concat('{"customer":',get_json_object(details,'$.customer'),',"order":',get_json_object(details,'$.order'),'"}') ELSE details END AS details FROM customer;