Json object selection - json

Question

I have a large JSON object returned from a Google Maps reverse geocode lookup.
in this object there are many possible location details returned.
Using jq how do I select the first returned location with a location_type "ROOFTOP" and get the formatted_address and place_id ?
in the below example the first entry has a location_type of "GEOMETRIC_CENTER"
The second entry has the "ROOFTOP" location_type. I want to ignore the entry with "GEOMETRIC_CENTER" and only return the fist entry with location_type "ROOFTOP"
Many thanks
Mike
[
{
"address_components": [
{
"long_name": "30",
"short_name": "30",
"types": [
"street_number"
]
},
{
"long_name": "Allée Jean de Lattre de Tassigny",
"short_name": "Allée Jean de Lattre de Tassigny",
"types": [
"route"
]
},
{
"long_name": "Montpellier",
"short_name": "Montpellier",
"types": [
"locality",
"political"
]
},
{
"long_name": "Hérault",
"short_name": "Hérault",
"types": [
"administrative_area_level_2",
"political"
]
},
{
"long_name": "Occitanie",
"short_name": "Occitanie",
"types": [
"administrative_area_level_1",
"political"
]
},
{
"long_name": "France",
"short_name": "FR",
"types": [
"country",
"political"
]
},
{
"long_name": "34000",
"short_name": "34000",
"types": [
"postal_code"
]
}
],
"formatted_address": "30 Allée Jean de Lattre de Tassigny, 34000 Montpellier, France",
"geometry": {
"bounds": {
"northeast": {
"lat": 43.6097932,
"lng": 3.8817559
},
"southwest": {
"lat": 43.6094097,
"lng": 3.881321299999999
}
},
"location": {
"lat": 43.6095516,
"lng": 3.881559199999999
},
"location_type": "GEOMETRIC_CENTER",
"viewport": {
"northeast": {
"lat": 43.6109504302915,
"lng": 3.882887580291503
},
"southwest": {
"lat": 43.6082524697085,
"lng": 3.880189619708498
}
}
},
"place_id": "ChIJ13k0paCvthIRcTgwBrisc10",
"types": [
"premise"
]
},
{
"address_components": [
{
"long_name": "8",
"short_name": "8",
"types": [
"street_number"
]
},
{
"long_name": "Rue Michelet",
"short_name": "Rue Michelet",
"types": [
"route"
]
},
{
"long_name": "Montpellier",
"short_name": "Montpellier",
"types": [
"locality",
"political"
]
},
{
"long_name": "Hérault",
"short_name": "Hérault",
"types": [
"administrative_area_level_2",
"political"
]
},
{
"long_name": "Occitanie",
"short_name": "Occitanie",
"types": [
"administrative_area_level_1",
"political"
]
},
{
"long_name": "France",
"short_name": "FR",
"types": [
"country",
"political"
]
},
{
"long_name": "34000",
"short_name": "34000",
"types": [
"postal_code"
]
}
],
"formatted_address": "8 Rue Michelet, 34000 Montpellier, France",
"geometry": {
"location": {
"lat": 43.60911189999999,
"lng": 3.8814264
},
"location_type": "ROOFTOP",
"viewport": {
"northeast": {
"lat": 43.61046088029149,
"lng": 3.882775380291502
},
"southwest": {
"lat": 43.60776291970849,
"lng": 3.880077419708498
}
}
},
"place_id": "ChIJ2UaGt6CvthIRJoNW7vS2Ibs",
"plus_code": {
"compound_code": "JV5J+JH Montpellier, France",
"global_code": "8FM5JV5J+JH"
},
"types": [
"street_address"
]
}
]

Answer 1

Rather than using .[0] to obtain the first object satisfying, it would be more efficient to use first/1. Here is a solution using first/1 and ..:
$ jq 'first(..|objects|select(.geometry.location_type == "ROOFTOP"))
| {formatted_address, place_id}' input.json
{
"formatted_address": "8 Rue Michelet, 34000 Montpellier, France",
"place_id": "ChIJ2UaGt6CvthIRJoNW7vS2Ibs"
}
Using firstin this way avoids having to compute the entire array before taking the first element.
Alternatively ...
Retrieving the first top-level object, without using ..:
first(.[] | select(.geometry.location_type=="ROOFTOP"))

Answer 2

Assuming the entries are enclosed in an array, you can use this jq filter that gives all objects having ROOFTOP as location:
jq '.[] | select(.geometry.location_type=="ROOFTOP")' file
If you want to select only the first one from all objects, use the following:
jq 'map(select(.geometry.location_type=="ROOFTOP"))|.[0]' file
And you only need need some of the values, pipe it another filter like this:
jq 'map(select(.geometry.location_type=="ROOFTOP"))|.[0]|.place_id, .formatted_address' file