SOME NOTES ON R FACILITIES FOR USE IN ASSIGNMENT 1


Debugging:

   options(warn=2)              Have warnings treated as errors (so they're
                                easier to debug)

   traceback()                  Right after an error, shows which function
                                it occured in

   options(error=dump.frames)   Allow use of the debugger after an error

   debugger()                   Start up the debugger, right after an error


Reading data from files:

   scan("file")                Good for reading a vector of numbers

   read.table("file",head=F)   Reads data with one case per line, returning
                               it as a "data frame"

   as.matrix(df)               Converts a data frame to a matrix.  Operations on
                               matrices can be much faster than on data frames

Sequences and grids:

   1:n                         Vector c(1,2,...,n-1,n).  Warning: using 1:0
                               doesn't give a zero-length vector!

   seq(a,b,length=n)           sequence of n numbers equally-spaced, starting
                               at a and ending at b

   rep(v,times=n)              Vector v repeated n times
   rep(v,each=n)               Vector formed by repeating each element of v
                               n times

Plotting:

   par(mfrow=c(1,2))           Put two plots on one page

   pdf("file.pdf",width=w,height=h)     Directs a plot to a PDF file, which
   plot(...)                            can then be viewed or printed with
   dev.off()                            acroread

   plot(x,y)                   Scatterplot of data in vectors x and y (which
                               must be the same length)

   contour(gx,gy,M)            Contour plot of data in the matrix M, with
                               the vectors gx and gy giving the coordinates
                               of rows and columns in M.

Matrices:

   as.vector(m)                 Useful for converting a 1-by-n or n-by-1 
                                matrix to a plain vector

   %*%                          Does a matrix-by-matrix multiply, or a 
                                matrix-by-vector or vector-by-matrix multiply

   M[i,]                        The vector consisting of row i of M

   M[-i,]                       The matrix consisting of all rows of M
                                EXCEPT row i (assuming i is positive).

   diag(v)                      Diagonal matrix with v on the diagonal; diag(n)
                                with n scalar is an n-by-n identity matrix

   cbind(...), rbind(...)       Put together matrices by concatenating vectors
                                or matrices horizontally or vertically

   solve(M)                     Inverse of M; solve(M,v) solves M %*% x = v

   t(M)                         Transpose


Programming:

   for (i in 1:n)               Loop with i taking on values 1, 2, ..., n
   { ...                        Warning!  Doesn't work if n is 0
   }

   if (..c..) { ..A.. }         If statement, does A if c is true, B otherwise,
   else { ..B.. }               can omit the "else" part