jq - extracting a subset of matching documents from a stream - json

Question

I have the following JSON message. This is a single complete message. There are so many messages like this in a single file. This json message was generated from a unformated json message using jq.
{
"header": {
"user": "baskar"
},
"requests": [
{
"first_name": "mike",
"last_name": "mat"
},
{
"first_name": "mike",
"last_name": "mat"
}
],
"check": [
"Y"
]
}
{
"header": {
"user": "baskar"
},
"message": {
"header": {
"user": "baskar"
},
"response": {
"resultsList": {
"result": [
{
"first_name": "mike1",
"last_name": "mat"
}
]
},
"errorMsg": null
}
}
}
I would like to do some filtering on this. For example, when i search for first_name, mike1, I should get the header and the matching request inside request. Also the matching result inside the Response message. So, the output is expected as follows for the search string mike1.
{
"header": {
"user": "baskar"
},
"requests": [
{
"first_name": "mike1",
"last_name": "mat"
}
],
"check": [
"Y"
]
}
{
"header": {
"user": "baskar"
},
"message": {
"header": {
"user": "baskar"
},
"response": {
"resultsList": {
"result": [
{
"first_name": "mike1",
"last_name": "mat"
}
]
},
"errorMsg": null
}
}
}
Basically, i want to filter out unmatched request inside the requests array and unmatched result inside the result array.
Currently, I use the following script to get the formatted json message from the unformatted json message log file.
sed -n "/<SEARCH_STRING>/ s/.*Service - //p" $1/test.log* | jq . > ~/result.log
Thanks,
Baskar.S

Answer 1

jq --arg key first_name \
--arg value mike1 \
'select(.message.response.resultsList.result[]?[$key]==$value) | .message' \
<in.json
...returns only message content where the result list contains at least one first_name of mike1.