Splitting nested arrays as separate entities - json

Question

I have some JSON data which contains attributes and some array elements. I would like to push a given set of fields into the array elements and then separate the arrays as separate entities.
Source data looks like this
[
{
"phones": [
{
"phone": "555-555-1234",
"type": "home"
},
{
"phone": "555-555-5678",
"type": "mobile"
}
],
"email": [
{
"email": "a#b.com",
"type": "work"
},
{
"email": "x#c.com",
"type": "home"
}
],
"name": "john doe",
"year": "2012",
"city": "cupertino",
"zip": "555004"
},
{
"phones": [
{
"phone": "555-666-1234",
"type": "home"
},
{
"phone": "555-666-5678",
"type": "mobile"
}
],
"email": [
{
"email": "a#b.com",
"type": "work"
},
{
"email": "x#c.com",
"type": "home"
}
],
"name": "jane doe",
"year": "2000",
"city": "los angeles",
"zip": "555004"
}
]
I expect a result like this
{
"person": [
{
"name": "john doe",
"year": "2012",
"city": "cupertino",
"zip": "555004"
},
{
"name": "jane doe",
"year": "2000",
"city": "los angeles",
"zip": "555004"
}
],
"phones": [
{
"name": "john doe",
"year": "2012",
"phone": "555-555-1234",
"type": "home"
},
{
"name": "john doe",
"year": "2012",
"phone": "555-555-5678",
"type": "mobile"
},
{
"name": "jane doe",
"year": "2000",
"phone": "555-666-1234",
"type": "home"
},
{
"name": "jane doe",
"year": "2000",
"phone": "555-666-5678",
"type": "mobile"
}
],
"email": [
{
"name": "john doe",
"year": "2012",
"email": "a#b.com",
"type": "work"
},
{
"name": "john doe",
"year": "2012",
"email": "x#c.com",
"type": "home"
},
{
"name": "jane doe",
"year": "2000",
"email": "a#b.com",
"type": "work"
},
{
"name": "jane doe",
"year": "2000",
"email": "x#c.com",
"type": "home"
}
]
}
I have been able to get the desired result, but I can't make it work in a generic way.
experiment on jqterm
The code below achieves the job, but I would like to pass the array of columns to be injected into the child arrays, the name of the primary result and an array containing the array field names.
["phones", "email"] as $children
| ["name", "year"] as $ids
|{person: map(with_entries(
. as $data | select($children|contains([$data.key])|not)
))}
+ {"phones": split_child($children[0];$ids)}
+ {"email": split_child($children[1];$ids)}

Answer 1

It's a lot more easier to achieve this using multiple reduces, like:
def split_data($parent; $ids; $arr_cols):
($arr_cols | map([.])) as $p
| reduce .[] as $in ({}; .[$parent] += [$in | delpaths($p)]
| (reduce $ids[] as $k ({}; . + {($k): $in[$k]}) as $s
| reduce $arr_cols[] as $k (.; .[$k] += [$in[$k][] + $s])
);
split_data("person"; ["name", "year"]; ["phones", "email"])

Answer 2

Here's a straightforward solution to the generic problem (it uses reduce only once, in a helper function). To understand it, it might be helpful to see it as an abstraction of this concrete solution:
{ person: [.[] | {name, year, city, zip} ]}
+ { phones: [.[] | ({name, year} + .phones[]) ]}
+ { email: [.[] | ({name, year} + .email[]) ]}
Helper function
Let's first define a helper function for constructing an object by selecting a set of keys:
def pick($ary):
. as $in
| reduce $ary[] as $k ({};
. + {($k): $in[$k]});
split_data
Here finally is the function that takes as arguments the $parent, $ids, and columns of interest. The main complication is ensuring that the supplemental keys ("city" and "zip") are dealt with in the proper order.
def split_data($parent; $ids; $arr_cols):
(.[0]|keys_unsorted - $arr_cols - $ids) as $extra
| { ($parent): [.[] | pick($ids + $extra)] }
+ ([$arr_cols[] as $k
| {($k): [.[] | pick($ids) + .[$k][]] }] | add) ;
The invocation:
split_data("person"; ["name", "year"]; ["phones", "email"])
produces the desired result.