Splitting Matrix into multiple couples of columns in R -

using r have 10 rows x 6 columns matrix. need split submatrices throuhg gruoping couples of columns no overlapping.

i.e. matrix has columns a,b,c,d,e,f , need extract 3 different matrices (or data.frames or whatever object within financial packages zoo or timeseries) formed columns ab, cd , ef.

ps: matrix contains financial data series , couple of columns has date columns , nav column

using dummy data (note must have dataframe otherwise r not allow hold date , numeric values in matrix [unless converted characters or raw numeric representations])

set.seed(42) df <- data.frame(a = sys.date() + 0:9, b = rnorm(10),                  c = sys.date() - 0:9, d = rnorm(10),                  e = sys.date() - 20:29, f = rnorm(10))  > head(df)                     b          c          d          e          f 1 2013-04-05  1.3709584 2013-04-05  1.3048697 2013-03-16 -0.3066386 2 2013-04-06 -0.5646982 2013-04-04  2.2866454 2013-03-15 -1.7813084 3 2013-04-07  0.3631284 2013-04-03 -1.3888607 2013-03-14 -0.1719174 4 2013-04-08  0.6328626 2013-04-02 -0.2787888 2013-03-13  1.2146747 5 2013-04-09  0.4042683 2013-04-01 -0.1333213 2013-03-12  1.8951935 6 2013-04-10 -0.1061245 2013-03-31  0.6359504 2013-03-11 -0.4304691 

one easy way form index columns want - here chose first column of each pair, 1, 3, 5, etc.

start <- seq(1, = 2, length = ncol(df) / 2) 

then, lapply on indices in start , select our data frame ith , ith + 1 columns i takes each index start in turn (df[i:(i+1)])

sdf <- lapply(start, function(i, df) df[i:(i+1)], df = df) 

which gives:

> sdf [[1]]                       b 1  2013-04-05  1.37095845 2  2013-04-06 -0.56469817 3  2013-04-07  0.36312841 4  2013-04-08  0.63286260 5  2013-04-09  0.40426832 6  2013-04-10 -0.10612452 7  2013-04-11  1.51152200 8  2013-04-12 -0.09465904 9  2013-04-13  2.01842371 10 2013-04-14 -0.06271410  [[2]]             c          d 1  2013-04-05  1.3048697 2  2013-04-04  2.2866454 ....  > str(sdf) list of 3  $ :'data.frame':   10 obs. of  2 variables:   ..$ a: date[1:10], format: "2013-04-05" "2013-04-06" ...   ..$ b: num [1:10] 1.371 -0.565 0.363 0.633 0.404 ...  $ :'data.frame':   10 obs. of  2 variables:   ..$ c: date[1:10], format: "2013-04-05" "2013-04-04" ...   ..$ d: num [1:10] 1.305 2.287 -1.389 -0.279 -0.133 ...  $ :'data.frame':   10 obs. of  2 variables:   ..$ e: date[1:10], format: "2013-03-16" "2013-03-15" ...   ..$ f: num [1:10] -0.307 -1.781 -0.172 1.215 1.895 ... 

an advantage of keeping sub-data frames in list can apply function or other operation sub-data frames using loop or tools lapply or sapply example.