jq - create new object based on existing objects - json

Question

This is my very first post, so please forgive me if it is not perfect.
I have json like below. It consists of two parts. The first part contains information about user (id, fullname and email) and the second part contains informations to which team user belongs. (id, team, role)
What I want to get an object which contains: id, fullName, emaila, team and role. I can do it but only when user belongs to one team. If user belongs to more than one team - I can't handle it.
Below my json:
[
{
"id": "user1",
"fullName": "User One",
"email": "user.one#my.mail.com"
},
{
"id": "user2",
"fullName": "User Two",
"email": "user.two#my.mail.com"
},
{
"id": "user1",
"team": "Team_A",
"role": "TEAM_MEMBER"
},
{
"id": "user1",
"team": "Team_B",
"role": "TEAM_ADMIN"
},
{
"id": "user2",
"team": "Team_B",
"role": "TEAM_ADMIN"
}
]
When I use: group_by(.id)[] | add
I get:
{
"id": "user1",
"fullName": "User One",
"email": "user.one#my.mail.com",
"team": "Team_B",
"role": "TEAM_ADMIN"
},
{
"id": "user2",
"fullName": "User Two",
"email": "user.two#my.mail.com",
"team": "Team_B",
"role": "TEAM_ADMIN"
}
and it is almost what I want to achieve. My goal is:
{
"id": "user1",
"fullName": "User One",
"email": "user.one#my.mail.com",
"team": "Team_A,
"role": "TEAM_MEMBER"
},
{
"id": "user1",
"fullName": "User One",
"email": "user.one#my.mail.com",
"team": "Team_B",
"role": "TEAM_ADMIN"
},
{
"id": "user2",
"fullName": "User Two",
"email": "user.two#my.mail.com",
"team": "Team_B",
"role": "TEAM_ADMIN"
}
I tried reduce also, but with no success.
I have made many attempts, but none of them gave the desired effect.
How is it possible using jq?
Thanks in advance,
krzyhon

Answer 1

It looks like you have a collection of bits of information about users. You can't just flatten them down by merging everything that has the same id, they need to be separate.
You have "user info" (full name and email) and "team info". You need to group by the id, then by the type, then distribute the "user info".
Here's one approach you could take.
# partition the data by "user/team type"
reduce .[] as $i ({}; if "fullName" | in($i) then .user += [$i] else .team += [$i] end)
# create a lookup of "user" data
| (.user | INDEX(.id)) as $user
# group the "team" objects by team
| .team | group_by(.team)
# merge corresponding "user info" with all team objects
| map(map(. + $user[.id]))
[
[
{
"id": "user1",
"team": "Team_A",
"role": "TEAM_MEMBER",
"fullName": "User One",
"email": "user.one#my.mail.com"
}
],
[
{
"id": "user1",
"team": "Team_B",
"role": "TEAM_ADMIN",
"fullName": "User One",
"email": "user.one#my.mail.com"
},
{
"id": "user2",
"team": "Team_B",
"role": "TEAM_ADMIN",
"fullName": "User Two",
"email": "user.two#my.mail.com"
}
]
]
jqplay
Here's another more concise solution, assuming there will only be one "user info" per id.
# group by id
group_by(.id) | map(
# for each group, partition by "type"
group_by(.fullName)
# create combinations of all the info and team objects and merge them
| combinations | add
)
jqplay

Answer 2

This solution uses a variant of #JeffMercado's approach based on combinations, but only uses group_by once (which means that for large datasets, it should be more efficient, since group_by is a relatively expensive operation).
The proposed solution here produces an array of the JSON objects you indicate you want, but if you want a stream of the JSON objects, simply omit the outer square brackets.
[group_by(.id)[]
| [map(select(has("team")|not)), map(select(has("team")))]
| combinations
| add]
More efficiently ...
To avoid using group_by or select at all, we could use the following generic variant of group_by:
def aggregate_by(s; f):
reduce s as $x (null; .[$x|f] += [$x]);
The solution can now be written as follows:
[ aggregate_by(.[]; .id)[]
| aggregate_by(.[]; .team == null | tostring)
| [.[]]
| combinations
| add ]