Parsing json with JQ : Where are the comma? - json

Question

I've this type of data :
{
"cidr" : "X.X.X.X/27",
"defaultGateway" : "X.X.X.X",
"full" : false,
"id" : "X.X.X.X",
"ipAddressTab" : [
{
"alias_domain" : null,
"alias_name" : null,
"description" : "This is the network address for X.X.X.X/27",
"dnr_rr" : null,
"dns_domain" : null,
"environnement" : null,
"fdqn" : null,
"hostname" : null,
"ip" : "X.X.X.X",
"requester" : null,
"status" : "reserved",
"type" : "network"
},
{
"alias_domain" : null,
"alias_name" : null,
"description" : "This is the default gateway address for X.X.X.X/27",
"dnr_rr" : null,
"dns_domain" : null,
"environnement" : null,
"fdqn" : null,
"hostname" : null,
"ip" : "X.X.X.X",
"requester" : null,
"status" : "reserved",
"type" : "gateway"
},
{
"alias_domain" : "toto.com",
"alias_name" : "",
"description" : "this is a test",
"dns_domain" : "",
"environnement" : "test",
"fdqn" : "XXX",
"hostname" : "XXX",
"ip" : "X.X.X.X",
"requester" : "XXX",
"status" : "allocated",
"type" : "VM"
},
{
"alias_domain" : "toto.com",
"alias_name" : "",
"description" : "this is a test",
"dns_domain" : "",
"environnement" : "test",
"fdqn" : "XXX",
"hostname" : "XXX",
"ip" : "X.X.X.X",
"requester" : "XXX",
"status" : "allocated",
"type" : "VM"
},
{
"ip" : "X.X.X.X",
"status" : "reserved"
},
{
"ip" : "X.X.X.X",
"status" : "reserved"
},
{
"ip" : "X.X.X.X",
"status" : "reserved"
},
{
"ip" : "X.X.X.X",
"status" : "reserved"
},
{
"alias_domain" : null,
"alias_name" : null,
"description" : "This is the broadcast address for X.X.X.X/27",
"dnr_rr" : null,
"dns_domain" : null,
"environnement" : null,
"fdqn" : null,
"hostname" : null,
"ip" : "X.X.X.X",
"requester" : null,
"status" : "reserved",
"type" : "broadcast"
}
]
}
I want to extract the IPs with "reserved" as status but without extractig the data for the gateway/network/broadcast IPs :
cat myfile | jq '.ipAddressTab[] | select(.status == "reserved") | select(.type != "network") | select(.type != "gateway") | select(.type != "broadcast")'
The output is :
{
"ip": "X.X.X.X",
"status": "reserved"
}
{
"ip": "X.X.X.X",
"status": "reserved"
}
{
"ip": "X.X.X.X",
"status": "reserved"
}
{
"ip": "X.X.X.X",
"status": "reserved"
}
But where are the commas between each " IP part " ?
On my file, each part in Braces are separated by a comma :
{
"ip" : "X.X.X.X",
"status" : "reserved"
}, <=
{
"ip" : "X.X.X.X",
"status" : "reserved"
}, <=
{
"ip" : "X.X.X.X",
"status" : "reserved"
}, <=
{
"ip" : "X.X.X.X",
"status" : "reserved"
}, <=
There is a way to keep them when I parse my file with jq ?
Thanks !

Answer 1

JQ's input and output are streams of JSON entities. The items in those streams are not comma separated. Some JQ filters may emit more than one entity for each entity they read in. Some filters may not emit any entity for some of the entities that they read in.
In your case, your input is a stream of one object, and your output is a stream of multiple objects. This is because each single array input to the [] filter results in a stream of all of the items in the output. It sounds like you instead want your output to consist of one array of objects. Here are two broad ways to do it.
1. Use map instead of []
cat myfile | jq '.ipAddressTab | map(select(.status == "reserved") | select(.type != "network") | select(.type != "gateway") | select(.type != "broadcast"))'
The parameter to map is another filter. It will be applied to each item in the array, and all the entities emitted by that filter will be gathered into one new array.
2. Use [ ... ] to gather the results of a filter into an array
cat myfile | jq '.ipAddressTab | [ .[] | select(.status == "reserved") | select(.type != "network") | select(.type != "gateway") | select(.type != "broadcast") ]'
It might seem like putting values between square brackets is just creating an array with one entry, but in fact what this syntax means is to evaluate the filter between the brackets, take the entire stream of output from that filter and make an array out of it.
In fact, this is exactly how map is itself defined. But it's useful to know, because it lets you see how you might use [ ... ] in other ways to capture a sequence of values.