squeue can produce JSON output with squeue --json.That could be processed with a script but you can do some quick hacks with jq.
Show just a few job fields
$ squeue --json | jq '.jobs[] | {job_id, user_name, account}'
{
"job_id": 66793003,
"user_name": "foo",
"account": "def-foo_cpu"
}
{
"job_id": 66844842,
"user_name": "bar",
"account": "def-bar_cpu"
}
...
List of all accounts currently running jobs, output in raw format
$ squeue --json | jq -r '.jobs | map(.account) | unique | .[]'
cc-debug_cpu
ctb-aaaa-ab_cpu
ctb-aaaa_cpu
ctb-aaaa_gpu
ctb-bbbb_cpu
ctb-cccc_gpu
ctb-dddd_gpu
def-eeee_cpu
How many unique accounts are running jobs?
$ squeue --json | jq '.jobs | map(.account) | unique | length'
251
How many jobs in each partition?
$ squeue --json | jq -r 'reduce .jobs[] as $job ({}; .[$job.partition] |= . + 1)'
{
"cpubase_bycore_b1,cpubackfill": 241,
"cpubase_bynode_b1,cpubackfill": 6,
"cpubase_bycore_b5": 2309,
"cpubase_bynode_b3,cpubackfill": 153,
"cpubase_bycore_b3,cpubackfill": 2,
"cpularge_bycore_b3,cpubackfill": 17,
"cpubase_bycore_b4": 2696,
"cpularge_bycore_b1,cpubackfill": 48,
"cpubackfill,cpubase_bynode_b1": 2,
"cpularge_bycore_b4": 81,
"cpubase_bynode_b4": 299,
"cpubase_bynode_b3": 29,
"cpubase_bynode_b5": 78,
"gpubase_bynode_b5": 18,
"gpubase_bygpu_b5": 125,
"cpubackfill,cpularge_bycore_b3": 1,
"cpularge_bycore_b5": 78,
"gpubase_bygpu_b4": 681,
"cpularge_bynode_b5": 39,
"cpubackfill,cpubase_bycore_b1": 1,
"cpubackfill,cpubase_bynode_b3": 14,
"gpubase_bynode_b4": 47,
"gpubase_bygpu_b2,gpubackfill": 2,
"cpularge_bynode_b4": 4,
"gpubase_bygpu_b3": 31,
"c-enamul": 1,
"cpubase_bycore_b3": 414,
"gpubase_bygpu_b3,gpubackfill": 8,
"cpubackfill": 34,
"gpubackfill": 7,
"cpularge_bycore_b3": 31,
"cpubase_bycore_b2": 17,
"cpularge_bynode_b3": 2,
"cpubase_bynode_b2": 36,
"cpularge_bycore_b2,cpubackfill": 4,
"cpularge_bycore_b2": 14,
"cpubase_bycore_b1": 101,
"cpubase_bynode_b1": 2,
"gpubase_bygpu_b2": 3,
"cpularge_bynode_b3,cpubackfill": 3,
"gpubase_bygpu_b1": 5,
"gpubase_bygpu_b1,gpubackfill": 1,
"cpularge_bycore_b1": 1,
"cpubase_bynode_b2,cpubackfill": 1
}
Show as an array of key/value pairs (for easier filtering)
$ squeue --json | jq -r 'reduce .jobs[] as $job ({}; .[$job.partition] |= . + 1) | to_entries'
[
{
"key": "cpubase_bycore_b1,cpubackfill",
"value": 241
},
{
"key": "cpubase_bynode_b1,cpubackfill",
"value": 6
},
{
"key": "cpubase_bycore_b5",
"value": 2309
},
...
Filter them to show only partitions with at least 100 jobs
(And we convert back to a JSON object after filtering for easier reading.)
$ squeue --json | jq -r 'reduce .jobs[] as $job ({}; .[$job.partition] |= . + 1) | to_entries | map(select(.value >= 100)) | from_entries'
{
"cpubase_bycore_b1,cpubackfill": 241,
"cpubase_bycore_b5": 2309,
"cpubase_bynode_b3,cpubackfill": 153,
"cpubase_bycore_b4": 2696,
"cpubase_bynode_b4": 299,
"gpubase_bygpu_b5": 125,
"gpubase_bygpu_b4": 681,
"cpubase_bycore_b3": 414,
"cpubase_bycore_b1": 101
}
Show the top five
$ squeue --json | jq -r 'reduce .jobs[] as $job ({}; .[$job.partition] |= . + 1) | to_entries | sort_by(.value) | reverse | .[:5] | from_entries'
{
"cpubase_bycore_b4": 2696,
"cpubase_bycore_b5": 2309,
"gpubase_bygpu_b4": 681,
"cpubase_bycore_b3": 414,
"cpubase_bynode_b4": 299
}