On 11/11/2011 07:50 PM, Paul Eggert wrote: > On 11/11/2011 09:55 AM, Pádraig Brady wrote: >> sort -b -k4.9,4.12 -k4.5b,4.7Mb -k4.2,4.3 -k4.14,4 access_log > Whew! That's pretty subtle. Do you observe > a significant performance improvement by avoiding > -n? I'm asking partly because the manual does it this way: > > sort -t ' ' -k 4.9n -k 4.5M -k 4.2n -k 4.14,4.21 > > I find this a bit easier to explain, but if it's significantly > slower we should probably fix the manual.
I tested that previously and noted only a small performance difference (I think). So I left the above as more appropriate for the manual. Testing again on my current hardware shows my version is 33% _slower_ than that in the manual! I guess explicitly checking the endpoints of each field is slower than checking implicit sentinels in the data. cheers, Pádraig.
