namespace about rapport HOT 8 CLOSED

Thanks, will do that soon for sure.

from rapport.

Hopefully fixed this with with my recent commit, where

• lattice stayed as a dependency - those functions should be really on global search path,
• same stands for plyr as reshape is looking for is.formula in global search path,
• other packages were moved to Imports from Depends will importing just the required functions - not all from there! So if new features would be added, pls update import fields too in rapport.R.

Please reopen if I did something lame here, had never ever dealt with Imports tag before.

from rapport.

Why are you using reshape and not reshape2? I think 'reshape' is depricated. That will most likely solve the namespace issue as well.

from rapport.

Thanks @jeroenooms for suggestion. I've just asked @aL3xa about this, but it seems that we had some issue with reshape2 before and it does depends on stringr which is a mortal enemy of @aL3xa who has a regexp in his pocket for each stringr example :)

Will report back here if we could sort things out, thanks again.

from rapport.

@daroczig, I checked out reshape2, and I can replace the old reshape stuff without a hustle... dunno what was the problem way back when... probably the notorious package that I seem to throw hate at... So, if you wanna change this, lemme know, and I'll fix the reshape stuff, and please don't give a bloody dime about my stringr rants. After all, y'know what they say: hater's gonna hate, and if stringr's gonna bring any good to package, I should STFU. So there, ploblem solved. =)

from rapport.

OK @jeroenooms , I did some testing/debugging with reshape2, and it seems to be faster, but with some new "features" that I really don't dig. The gist of rp.desc is to accept list of aggregation functions that are to be applied against the data. But dcast is apparently not in love with each and I get some strange stuff there... e.g. at some point, while debugging, I've received, for the first time in my R coding history, a segfault:

Browse[2]> dcast(m, f, fun.aggregate = each(fn), margins = margins, subset = subset, drop = FALSE)

 *** caught segfault ***
address 0x2000004a, cause 'memory not mapped'

Traceback:
 1: .Call("split_indices", index, group, as.integer(n))
 2: split_indices(seq_along(.value), .group, .n)
 3: vaggregate(.value = value, .group = overall, .fun = fun.aggregate,     ..., .default = fill, .n = n)
 4: cast(data, formula, fun.aggregate, ..., subset = subset, fill = fill,     drop = drop, value.var = value.var)
 5: dcast(m, f, fun.aggregate = each(fn), margins = margins, subset = subset,     drop = FALSE)
 6: rp.desc("age", "gender", mean, ius2008)

Possible actions:
1: abort (with core dump, if enabled)
2: normal R exit
3: exit R without saving workspace
4: exit R saving workspace

I'll rant about this on SO or R mailing list, but I'd rather stick with reshape. It works! =)

from rapport.

Wow that does look pretty nasty :) I read your SO post; honestly I wasn't aware that a lot of the functionality was stripped (I assumed @hadley just bumped to reshape2 because of some backward incompatible changes). Unfortunately I think he doesn't seem to be actively maintaining the old reshape anymore and putting his efforts in the next generation, so something to keep in the back of your mind.

from rapport.

I've just realised that reshape is actually reshape2 repo. Will report issue there. =)

from rapport.