CSV to JSON and add title - json

Question

I have a csv document:
{
"epsilon_id": 194029423,
"weather": "cloudy",
"temperature": 27
},
{
"epsilon_id": 932856192,
"weather": "sunny",
"temperature": 31
}
I was wondering if there was a tool to make it into valid json where the field epsilon_id is the title for the data.
ex:
{
194029423: {
"weather": "cloudy",
"temperature": 27
},
932856192: {
"weather": "sunny",
"temperature": 31
}
}
I would prefer it to be a program (in whatever language) that I can run because I have 1,000 entries in my test sample and I will have tens of thousands in my final copy.
Any help would be much appreciated!

Answer 1

You are looking at JSON transformation, and ofcourse can be achieved with a custom programming. I can explain you how you can achieve this in Java, but functionally its gonna be the same for any programming of your choice.
Your input json will look like this:
[{
"epsilon_id": 194029423,
"weather": "cloudy",
"temperature": 27
},
{
"epsilon_id": 932856192,
"weather": "sunny",
"temperature": 31
}]
When you parse in java using popular Jackson library, you will get list of object for below class:
class Input
{
#JsonProperty(access = Access.WRITE_ONLY)
String epsilon_id,
String weather,
int temperature
}
Then you create a map object Map<Integer, Input>, populate data like below:
Map<Integer, Input> map = new HashMap<>();
for(Input obj : listOfInputs){
map.put(obj.epsilon_id, obj)
};
Serialize your result map using Jackson again to get your desired output format:
{
194029423: {
"weather": "cloudy",
"temperature": 27
},
932856192: {
"weather": "sunny",
"temperature": 31
}
}
If you are not very familiar with Java & Jackson JSON parsing, I found this tutorial with code sample, which will give you headstart.

Answer 2

import csv, json, os
# rename this file or pass it in as process.argv[2]
# then pipe output into another file. or
with open("./foo.csv") as f:
output = {}
for line in csv.DictReader(f):
key = line.pop("epsilon_id")
if output.has_key(key):
print("Duplicate Id -> {} ".format(key))
output[key] = line
# then pipe this output into another file.
print(json.dumps(output, indent=2))
# or write a file
with open("/tmp/foo.json",'w') as f:
json.dump(output, f)
Python makes it pretty easy : this will detect all types from a csv file.
demo : https://repl.it/#markboyle/UsableIcyFeed