read.pdb - maxlines

[Lars Skjærven]

why not set default value of maxlines to -1 ?

[Barry Grant]

Because it could be very very slow with big structures and I would rather not leave a user hanging waiting for a prompt to come back. Also I want to encourage folks to use netcdf for multimodel files.

There is a danger with having a maxlines set as it is currently is that a user might continue with a truncated pdb object without noticing the returned warning message. So perhaps a warning would be better placed if the file is over a certain number of lines but the function should continue on with the maxlines=-1 setting and require a Ctrl-C?

[Lars Skjærven]

I would prefer maxlines=-1. It's not only multi model files. e.g. groel (pdb 1svt: 63k lines). also, must functions, e.g. read.table and readLines uses maxlines (or nrows) = -1.

[Barry Grant]

OK Lars, feel free to make the change when its convenient. Cheers, Barry

[Lars Skjærven]

241k lines runs in roughly 20 sec. Ctrl-C works.

> raw = readLines("ho.pdb")
> length(raw)
[1] 241514

> system.time(read.pdb("ho.pdb", maxlines=-1))
   user  system elapsed 
 21.635   0.068  21.771 

> system.time(read.pdb("ho.pdb", maxlines=-1))
^C
Timing stopped at: 10.762 0.016 10.811 

[Barry Grant]

Good to know we are not as slow as I thought!