So apparently it runs the query against an in-memory sqlite instance. I'm curious what the rough time overhead for that is (compared to, say, the overhead of calling out to grep). Does it feel noticeably slow to run a query on ps output?
And the other thing I'm wondering, can this drop you into a sqlite command line? I.e.,
» ps aux -> select -i
sql> SELECT ...
It took 0.2 seconds to run the command on my relatively modest 4 yr old MBP.
Running a similar pipeline (below) took 0.4 seconds. But there are optimisations that can be made to the command line:
time {ps aux -> awk {print $1} -> sort -> uniq -c -> sort}
["Name", "Session", "Score", "Completed"]
["Gilbert", "2013", 24, true]
["Alexa", "2013", 29, true]
["May", "2012B", 14, false]
["Deloise", "2012A", 19, true]
The long term plan for this feature I've written is to make it streamable. In fact I do already have some draft code written to do that. But I figured I'd get an early draft of this out there to gauge peoples thoughts.
> And the other thing I'm wondering, can this drop you into a sqlite command line?
Currently no. And it's not something I'm looking to add support for because there's already a lot of decent SQL tools out there. But I was thinking of adding an option to write the database to disk as a standard .db file and you can then use your favourite interactive sqlite3 tool to query the data.
Iirc, sort still needs to load the input in memory to be able to sort correctly, I presume. Uniq indeed works on the stream directly.