Merging two files using jq - json

Question

I've got two files.
First:
[
{ "person1": [] },
{ "person2": [] }
]
Second:
[
{
"person2": { "attribute1": "wer", "attribute2": "sdf" }
},
{
"person2": { "attribute1": "ert", "attribute2": "dfg" }
},
{
"person2": { "attribute1": "rty", "attribute2": "fgh" }
},
{
"person3": { "attribute1": "tyu", "attribute2": "ghj" }
},
{
"person1": { "attribute1": "yui", "attribute2": "hjk" }
}
]
I try to merge them, using jq. For each person from the first file (in the second file migth be more persons, which should be ignored) create list of it's attributes. So an output should look something like this:
[
{
"person1":
[
{ "attribute1": "yui", "attribute2": "hjk" }
]
},
{
"person2":
[
{ "attribute1": "wer", "attribute2": "sdf" },
{ "attribute1": "ert", "attribute2": "dfg" },
{ "attribute1": "rty", "attribute2": "fgh" }
]
}
]
I tried different options, but I can't achieve expected result.

Answer 1

jq 'reduce (input[]|to_entries[]) as $e (add;
if has($e.key) then .[$e.key] += [$e.value] else . end
) | [keys_unsorted[] as $k|{($k): .[$k]}]' file1 file2
online demo

Answer 2

The following solution is focused on efficiency, and also suggests an alternative format for the output:
jq program (merge.jq)
add as $dict
# aggregation:
| (reduce input[] as $record ({};
($record|keys_unsorted[0]) as $person
| if $dict[$person] then .[$person] += [$record[$person]] else . end )) as $answer
# re-arrangement
| reduce ($dict|keys_unsorted[]) as $person ([]; . + [ {($person): $answer[$person] } ] )
Invocation
jq -f merge.jq first.json second.json