R: my cookbook

라벨: Informatics
이메일로 전송BlogThis!X에 공유Facebook에서 공유

{My R Cookbook}

ToC

• transform() column class/mode
• reshape table
• add a new column with a variable name 
• flexible R ps/pdf font handling
• Match
• Direct data input
• Sum() by a reference column
• Not consistent column numbers in table?
• String as variable name
• Memory-efficient coding
• Data-type issue
• split strings in a vector and convert it to a data.frame
• Installation 
• read table with variable column numbers

# transform() column class/mode

 transform(d, fake_char = as.numeric(fake_char), 
               char_fac = as.numeric(char_fac))

# Reshape table

http://stackoverflow.com/questions/9617348/reshape-three-column-data-frame-to-matrix

[FROM]

x a 1 
x b 2 
x c 3 
y a 3 
y b 3 
y c 2 

[TO]

    a    b    c
x   1    2    3
y   3    3    2

[HOW TO]?

tmp <- data.frame(x=gl(2,3, labels=letters[24:25]),
                  y=gl(3,1,6, labels=letters[1:3]), 
                  z=c(1,2,3,3,3,2))

[Using reshape2:]

library(reshape2)
acast(tmp, x~y, value.var="z")

[Using matrix indexing:]

with(tmp, {
  out <- matrix(nrow=nlevels(x), ncol=nlevels(y),
                dimnames=list(levels(x), levels(y)))
  out[cbind(x, y)] <- z
  out
})

[Using xtabs:]

xtabs(z~x+y, data=tmp)

[Using reshape()]

> reshape(tmp, idvar="x", timevar="y", direction="wide")
  x z.a z.b z.c
1 x   1   2   3
4 y   3   3   2

# Add a new column with a variable name


x = cbind( a=1:2, b=3:4)
x = data.frame(x)   # should be a data frame
new = 'c'
x[[ new]] = 5:6          # this works only with data frame
x


#  Flexible R ps/pdf fonts handling

http://blog.revolutionanalytics.com/2012/09/how-to-use-your-favorite-fonts-in-r-charts.html

library(extrafont)

library(extrafont)
font_import()

fonts()

names(odfFonts())

 
# Match

which( 20 == c(11,22,20))
[1] 3

c(11,20,30) %in% c(10,33,20)
[1] FALSE TRUE FALSE
 
match( c(10,20,30), c(11,33,20) )
[1] NA 3 NA




# Direct data input

dat <- read.table(header = TRUE,  text = "SNP Geno    Allele
marker1 G1  AA
marker2 G1  TT
marker3 G1  TT
marker1 G2  CC
marker2 G2  AA
marker3 G2  TT
marker1 G3  GG
marker2 G3  AA
marker3 G3  TT")

# Sum column values by a reference column

tappy( t$vak, t$ref, FUC=sum)  # only one column


or

# for multiple columns
aggreagte( t[,3:10], by=list(symbol=$symbol), FUN=sum)



# Version check

str(.Platform)

version

# Not the consistent column in table?

gmt = "c2.all.v4.0.symbols.gmt"
gmt = "c2-kegg-biocarta-reactome.gmt"
fieldCount = count.fields(gmt, sep = "\t")
msigdb = read.csv( file=gmt, sep="\t", header=F, stringsAsFactor=F, fill=T, col.names=1:max(fieldCount))

## for fast computation, use matrix
msigdb = as.matrix( msigdb )


# Check classes of columns  | reading table faster

x = read.table( file="xxx", nrows=10000)  
 
sapply( x, class )
read.table( file="xxx", colClasses = sapply(x, class) ) 


# String as variable name

assign("x", 5)

i=3
assign( paste("a.", i, sep=""), 6 )
get( paste("a.",i,sep="")


# Matrix

> x = matrix(1:6,2,3)

> x
     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6

> x[[3]]
[1] 3

> x[1,2,drop=F]
     [,1]
[1,]    3

> x[1, , drop=F]
     [,1] [,2] [,3]
[1,]    1    3    5

# List

x = list( foo=1:4, bar=0.6 )

x[1]

x[1]
$foo
[1] 1 2 3 4

x[[1]]

[1] 1 2 3 4

> x$bar
[1] 0.6


> x[["bar"]]
[1] 0.6


> x["bar"]
$bar
[1] 0.6
> x= list( aadsffs = 1:5, afdfsaf=1:6)

> x$aa
[1] 1 2 3 4 5
> x[["aa", exact=F]]
[1] 1 2 3 4 5

> complete.cases( vector1, vector2 ) # remove NA

# Memory-efficient

Source: http://www.r-bloggers.com/faster-for-loops-in-r/

slow!

x = c()
system.time(
  for(i in 1:40000){
    x = c(x,i) #here i is combined with previous contents of x
  }
)

   user  system elapsed 
   1.54    0.00    1.53 

Much faster trick!

x = rep( NA, 50000 )
x = NULL
system.time(
  for(i in 1:40000){
    x[i] = i #changing value of particular element of x
  }
)

  user  system elapsed 
   0.57    0.00    0.56 

EVEN faster trick!

x = rep( NA, 50000 )
x = rep(NA, 1000000)
system.time(
  for(i in 1:40000){
    x[i] = i #changing value of particular element of x
  }
)
x[!is.na(x)]

   user  system elapsed 
   0.06    0.00    0.06 

Warning: Data types! as.numeric()

http://stackoverflow.com/questions/2288485/how-to-convert-a-data-frame-column-to-numeric-type

d = data.frame(char = letters[1:5], 
                fake_char = as.character(1:5), 
                fac = factor(1:5), 
                char_fac = factor(letters[1:5]), 
                num = 1:5, stringsAsFactors = FALSE)

sapply(d, mode)
sapply(d, class)
transform(d, fake_char = as.numeric(fake_char), 
               char_fac = as.numeric(char_fac))

split strings in a vector and convert it to a data.frame

string = c("xy_100_ab", "xy_101_ab","xy_102_ab","xy_103_ab")

out = data.matrix( do.call( rbind, strsplit( string, '_' ) ) )
## if you use dataframe, your number will be factor, not character.

names(out) = paste('column',1:3,sep="")

Concatenate vectors

paste( v, collapse="" )



Get all pkgs list and re-install

• In old R
• IP = as.data.frame(installed.packages()) 
• MyPkgs <- subset(IP, !Priority %in% c("base", "recommended"), select = c(Package, Bundle)) 
• Then in a new version R,
• install.packages(MyPkgs$Package, dependencies = TRUE)

#--run in the old version of R
setwd("C:/Temp/")
packages <- installed.packages()[,"Package"]
save(packages, file="Rpackages") 
INSTALL NEW R VERSION
#--run in the new version
setwd("C:/Temp/")
load("Rpackages")
for (p in setdiff(packages, installed.packages()[,"Package"]))
install.packages(p) 


Installation Error

• custome build: ./configure --enable-R-shlib
• "configure: error: --with-readline=yes (default) and headers/libs are not available".
• SOLUTION: yum install readline-devel

Basic

1. How to delete previously saved workspace restored?
• unlink(".RData")

Text processing

1. Why does read.table() stop reading table while read.delim() works just fine? 
• http://stackoverflow.com/questions/3016333/r-why-does-read-table-stop-reading-a-file

How to get the only first item from a list: do not use "unlist" problem with

a = c(1:10)  

b = LETTERS[seq( from = 1, to = 10 )] x = list( a, b )  # or x = list( "a"=a, "b"=b ) names( x ) = c( "a", "b" )

## unlist lapply(x, FUN="[", 1); 

lapply(x, FUN="[", 1);
unlist( lapply(x, FUN="[", 1) , use.names=F) 

sapply(x, FUN=length); 

http://www.mayin.org/ajayshah/KB/R/index.html



Read a table w/ variable column number

fieldCount = count.fields("table", sep = "\t")

read.csv( file="table", sep="\t", header=F, stringsAsFactor=F, fill=T, col.names=1:max(fieldCount))

R by example

Basics

Vectors, lists, matrices, data frames	Sorting	Prices and returns
Writing functions	Amazing R vector notation	Amazing R indexing notation
Making latex tabular objects	Associative arrays / hashes	Matrix notation (portfolio computations in financial economics)
Handling missing data

Reading files

Reading a file with a few columns of numbers, and look at what is there.	Reading a file involving dates	Reading in a file made by CMIE'sBusiness Beacon program
Reading and writing both ascii files and binary files. Also, measure speed of these.	Sending an R data object to someone else, either in email or as a binary file.	Make a "zoo" object, for handling time-series data.
Exporting and importing data.	Reading .gz .bz2 files and URLs	Directly reading Microsoft Excel files

Graphs

A grid of multiple pictures on one screen	Making PDF files that go into books/papers	A histogram with tails in red
z=f(x,y) using contour lines and colours	Show recessions using filled colour in a macro time-series plot	Plotting two series on one graph, one with a left y axis and another with a right y axis.

Probability and statistics

Tables, joint and marginal distributions	`Moving window' standard deviation	Quartiles/deciles tables/graphs. [Requires this data file.]
Distribution of sample mean and sample median	The bootstrap [Notes on boot()]	Doing MLE with your own likelihood function [Notes on MLE]
The strange Cauchy distribution	An example of simulation-based inference	Four standard operations with standard distributions
Two CDFs and a two-sample Kolmogorov-Smirnoff test	Simulation to measure size and power of a test

Regression

Doing OLS	Dummy variables in regression	Generate latex tables of OLS results
`Least squares dummy variable' (LSDV) or `fixed effects' model	Estimate beta of Sun Microsystems using data from Yahoo finance : Elaborate version, Terse version.	Nonlinear regression
Standard tests	Using orthogonal polynomials	A function that takes a model specification as an argument

Time-series analysis

ARMA estimation, diagnostics, forecasting

이메일로 전송BlogThis!X에 공유Facebook에서 공유

라벨: Informatics