you could try aggregate(), e.g.,
my.df <- data.frame(pathway = c(rep("pw.A", 2), rep("pw.B", 3),
rep("pw.C", 1)),
cond.one = c(0.5, NA, 0.4, NA, NA, NA),
cond.two = c(NA, 0.6, NA, 0.9, NA, 0.2),
cond.three = c(NA, NA, NA, NA, 0.1, NA))
aggregate(my.df[-1], my.df['pathway'], sum, na.rm = TRUE)
or
sum. <- function(x) if (all(is.na(x))) NA else sum(x, na.rm = TRUE)
aggregate(my.df[-1], my.df['pathway'], sum.)
I hope it helps.
Best,
Dimitris
On 5/7/2012 11:50 AM, Karl Brand wrote:
Esteemed UseRs,
This must be embarrassingly trivial to achieve with e.g., melt() and
cast(): deduplicating records ("pw.X" in example) for a given set of
responses ("cond.Y" in example).
Hopefully the runnable example shows clearly what i have and what i'm
trying to convert it to. But i'm just not getting it, ?cast that is! So
i'd really appreciate some ones patience to clarify this, using the
reshape package, or any other approach.
With sincere thanks in advance,
Karl
## Runnable example
## The data.frame i have:
library("reshape")
my.df <- data.frame(pathway = c(rep("pw.A", 2), rep("pw.B", 3),
rep("pw.C", 1)),
cond.one = c(0.5, NA, 0.4, NA, NA, NA),
cond.two = c(NA, 0.6, NA, 0.9, NA, 0.2),
cond.three = c(NA, NA, NA, NA, 0.1, NA))
my.df
## The data fram i want:
wanted.df <- data.frame(pathway = c("pw.A", "pw.B", "pw.C"),
cond.one = c(0.5, 0.4, NA),
cond.two = c(0.6, 0.9, 0.2),
cond.three = c(NA, 0.1, NA))
wanted.df
