# [R-SIG-Finance] colMeans(x) <> colStats(x, mean)

Mercurio Danilo 1850 SPI Danilo.Mercurio at Sparinvest.com
Fri Jan 21 15:42:17 CET 2011

```Hi!

first I would like to say that I am an enthusiastic user of the Rmetrics libraries and I am very greatful to all the developpers.
I would really appreciate  a few comments on the following problem though. As the example below shows the function "colMeans" (package base) treats Nas
differently from the function "colStats" (package timeSeires). In order to have the same tretment of Nas I could not find a better solution as to write a new function: "colSdsNAs" instead of "colSds". I need this because I want to use the function im "applySeries" and my data is similar to the "x" in the example.

Is this the only way or I can use the functions "colStats", "colSds" in some way that suits my problem?

-------
require(timeSeries)
require(fAssets)

a = c(rep(NA, 5), 6:12)
b = c(rep(NA, 3), 4:12)
x = cbind(a, b)

x = as.timeSeries(x)
dt0  = as.timeDate("20000101")
dt1 =  as.timeDate("20001201")
tStamps = timeSequence(from = dt0, to = dt1, by = "1 month")
time(x) = tStamps
colMeans(x)
colSds(x)
colStats(x, mean)
#
colMeans(x, na.rm = TRUE)
colSds(x, na.rm = TRUE)

by = periods(time(x), "3m", "1m")
applySeries(x, from = by\$from, to = by\$to, FUN = "colMeans")
applySeries(x, from = by\$from, to = by\$to, FUN = "colSds")

colSdsNAs <- function(x) apply(x, 2, "sd")
applySeries(x, from = by\$from, to = by\$to, FUN = "colSdsNAs")
------

Dr. Danilo Mercurio
ERSTE-SPARINVEST Kapitalanlagegesellschaft m.b.H
Quantitative Analyst, Global Strategies and Research

AT00951850 Global Strategies & Research
Sitz des Unternehmens: 1010 Wien, Habsburgergasse 1a
Tel.:   +43 (0) 50100 - 19957
Fax.:  +43 (0) 50100 9 - 19957
Mob.: +43 (0) 50100 6 - 19957

www.sparinvest.com
Handelsgericht Wien - FN 81876 g

ERSTE-SPARINVEST Kapitalanlagegesellschaft m.b.H. - Member of Erste Group

```