[ESS] Does R fix() or edit() work for you these days?]
Kasper Daniel Hansen
khansen at stat.Berkeley.EDU
Mon Sep 10 19:25:18 CEST 2007
On Sep 10, 2007, at 9:23 AM, 8rino-Luca Pantani wrote:
> Martin Maechler ha scritto:
>> My opinion still applies:
>> Using fix() or edit() is "wrong in principle"
>> when working with ESS.
> From my non-programmer-non-statistician I sincerely trust you,
> but please, could you be so kind to spend some time in briefly
> explaining me what's wrong with fix/edit under ESS ?
> I mostly use fix/edit only to see values in large dataframes, when I'm
> enough lazy not to use indexes.
> For such a use, which are the "right" alternatives, other than lines
> like df.dataframe[101:107, 2] or df.dataframe[df.dataframe
> $FACTOR=="foo", ]
Martin is advocating "The source is real". The short explanation is
that every R object you have, should have some _code_ creating that
object. However when someone uses fix to edit a function, suddenly
whatever source code (s)he has, is obsolete.
You are describing a situation where you just use "fix" to _look_ at
a dataframe, _not_ modify it. And that would be entirely within
Martin's paradigm. However, as soon as you modify the dataframe, you
do not have code creating that object anymore.
> BTW Which are your favourite way of naming objects ?
> In other words, for a dataframe, do you prefer df.foo. foo.df,
> (plain)foo or what else ?
> Are there some rules to avoid/reduce error in working with XEmacs/
> ESS/R ?
> I mean rules such as naming objects in dataframe always in maiuscule,
> vectors always in minuscule and things like that.
There are no such rules. The only approximate rule which many people
kind of uses is that names with "." in them refers to S3-methods/
objects/classes and names with camelCase like dataFrame refers to S4
> ESS-help at stat.math.ethz.ch mailing list
More information about the ESS-help