Return to Answer

Having settled on bash + awk, the OP approach makes perfect sense. But it is tedious. If written in python we could "import typer", write a function signature, and get {parsing, diagnostics, --help} "for free".

The fundamental issue is we chose an inappropriate calendar representation. Yes, we could finesse with subkeys for sorting. But it would be far more sensible to parse out something like a dt.datetime or even an ISO string.

Having settled on bash + awk, the OP approach makes perfect sense. But it is tedious. If written in python we could "import typer", write a one-line function signature, and get {parsing, diagnostics, --help} "for free".

The fundamental issue is we chose an inappropriate calendar representation. Yes, we could finesse it with subkeys for sorting. But it would be far more sensible to parse out something like a dt.datetime or even an ISO string.

Consider this example. We run against {April, May}. Machine Bob was added in May, and we see a single record for that. So count is 1, Bob simply did not exist in April. And then suppose Bob is decommissioned in November.

Consider this example. We run against {April, May}. Machine Bob was added in May, and we see a single record for that. So count is 1, Bob simply did not exist in April. And then suppose Bob is decommissioned in November, with similar effect.

Consider this example. We run against {April, May}. Machine Bob was added in May, and we see a single record for that. So count is 1. Bob simply did not exist in April. And then suppose Bob is decommissioned in November.

Consider this example. We run against {April, May}. Machine Bob was added in May, and we see a single record for that. So count is 1, Bob simply did not exist in April. And then suppose Bob is decommissioned in November.