Chain select and max_by on json doc with jq in bash - json

Question

I want to get the SnapshotIdentifier of the snapshot with the maximum SnapshotCreateTime, and filter it by ClusterIdentifier. Here is the command I'm using:
aws redshift describe-cluster-snapshots --region us-west-2 |
jq -r '.Snapshots[]
| select(.ClusterIdentifier == "dev-cluster")
| max_by(.SnapshotCreateTime)
| .SnapshotIdentifier '
Here is the json
{
"Snapshots": [
{
"EstimatedSecondsToCompletion": 0,
"OwnerAccount": "45645641155",
"CurrentBackupRateInMegaBytesPerSecond": 6.2857,
"ActualIncrementalBackupSizeInMegaBytes": 22.0,
"NumberOfNodes": 3,
"Status": "available",
"VpcId": "myvpc",
"ClusterVersion": "1.0",
"Tags": [],
"MasterUsername": "ayxbizops",
"TotalBackupSizeInMegaBytes": 192959.0,
"DBName": "dev",
"BackupProgressInMegaBytes": 22.0,
"ClusterCreateTime": "2016-09-06T15:56:08.170Z",
"RestorableNodeTypes": [
"dc1.large"
],
"EncryptedWithHSM": false,
"ClusterIdentifier": "dev-cluster",
"SnapshotCreateTime": "2016-09-06T16:00:25.595Z",
"AvailabilityZone": "us-west-2c",
"NodeType": "dc1.large",
"Encrypted": false,
"ElapsedTimeInSeconds": 3,
"SnapshotType": "manual",
"Port": 5439,
"SnapshotIdentifier": "thismorning"
}
]
}

Answer 1

max_by expects an array as input. Thus the following variant of your filter would work:
[.Snapshots[] | select(.ClusterIdentifier == "dev-cluster")]
| max_by(.SnapshotCreateTime)
| .SnapshotIdentifier
Based on your verbal description, it would seem you want to run max_by before select:
.Snapshots
| max_by(.SnapshotCreateTime)
| select(.ClusterIdentifier == "dev-cluster")
| .SnapshotIdentifier
If there is possibly more than one maximal object, you might want to use maximal_by rather than max_by:
def maximal_by(f):
(map(f) | max) as $mx
| .[] | select(f == $mx);