Slides and script

Author: jonsan21

3.Loops and Flow Control/Loops and Flow Control.R

@@ -0,0 +1,94 @@
+#Loops in R
+
+1:3
+for(i in 1:3) {
+  cat(i, "+", i, "=", i+i, "\n")
+}
+d <- data.frame(a = 1:2, b=2:3)
+
+for(x in d) {
+  cat("Column sum:", sum(x), "\n")
+}
+
+#Flow Control: if and if else statements
+
+for(i in 1:3){
+  if (i==2) cat("This index is even:","\n")
+  cat(i,"\n")
+} 
+
+for(i in 1:3){
+  if (i==2) cat("The index is 2","\n") else 
+    cat("The index is not 2","\n")
+} 
+
+#While loop example
+
+
+set.seed(886)
+
+k<-0 # number of big parts (>2)
+y<-abs(rnorm(1000)) # simulated part size
+i<-0 # index of parts
+# loop: 
+while(k<3 & i<1000){
+  i<-i+1
+  temp<-y[i]
+  k<-k+(temp>2)
+}
+rm(temp)
+i
+
+#Repeat loop example}
+
+set.seed(371)
+eye.colors<-c("brown","blue","green","yellow","grey")
+eyecolor<-data.frame(personId=1:100,color=
+                       sample(eye.colors,100,rep=T))
+i<-0
+list.of.ids<-numeric(0) # patient ID list
+#loop:
+repeat {
+  i<-i+1
+  if(eyecolor$color[i]=="yellow" | 
+     eyecolor$color[i]=="blue") next
+  list.of.ids<-c(list.of.ids,eyecolor$personId[i])
+  if(i==100 | length(list.of.ids)==20) break
+}
+list.of.ids
+
+#Loops and run times
+
+set.seed(853)
+y<-matrix(rnorm(1000000),nrow=1000)
+z<-0*y
+time1<-as.numeric(Sys.time())
+#loop:
+for(i in 1:1000){
+  for(j in 1:1000){
+    z[i,j]<-y[i,j]^2
+  }
+}
+time2<-as.numeric(Sys.time())
+
+# using object form in R:
+z<-y^2
+time3<-as.numeric(Sys.time())
+
+# run time increase factor:
+(time2-time1)/(time3-time2)
+
+#lapply() and sapply()
+
+set.seed(777)
+
+my.data<-data.frame(data1=rnorm(10),data2=rnorm(10),data3=rnorm(10))
+lapply(my.data,sum)
+sapply(my.data,sum)
+
+#lapply() and sapply() examples
+
+A<-matrix(1:9,nrow=3);B<-matrix(1:16,nrow=4);C=matrix(1:8,nrow=4)
+my.list<-list(A=A,B=B,C=C); my.list
+
+lapply(my.list,"[",,2)

4.Working with Vectors and Matrices/Working with Vectors and Matrices.R

@@ -0,0 +1,74 @@
+#Constructing vectors
+
+x<-9:17
+x
+
+y<-seq(0,1,length=11)
+y
+
+z<-rep(1:2, 5)
+z
+
+xz10<-c(x,z,10)
+xz10
+
+#Constructing matrices
+
+A<-rbind(1:3, c(1,1,2))
+A
+
+B<-cbind(1:3, c(1,1,2))
+B
+
+C<-matrix(c(1,0,0,1,1,0,1,1,1), nrow=3, ncol=3)
+C
+
+#Index and logical index
+
+x<- (-5):5
+x
+
+x[4:8]
+
+x[-c(1:3,9:11)]
+
+index<-abs(x)<3
+index 
+x[index]
+
+A<-matrix((-4):5, nrow=2, ncol=5)
+A
+A[A<0]
+
+A[A<0]<-0
+A
+
+A[2,]
+
+A[,c(2,4)]  
+
+#Properties of vectors and matrices
+
+A<-matrix(rep(c(TRUE,FALSE),2),nrow=2)
+B<-rnorm(4)
+C<-matrix(LETTERS[1:9],nrow=3)
+A;B;C
+mode(A);mode(B);mode(C)
+
+x<-matrix(c(NA,2:12),ncol=3)
+x
+length(x[1,])
+length(x)
+
+dim(x); dim(x)[2]
+
+#Naming rows and columns in a matrix
+
+x<-matrix(rnorm(12),nrow=4)
+x
+dimnames(x)[[2]]<-paste("data",1:3,sep="")
+dimnames(x)[[1]]<-paste("obs",1:4,sep="")
+x
+
+rownames(x)
+colnames(x)

5.Reading in Data/Reading in Data.R

@@ -0,0 +1,43 @@
+#Reading data from a text file
+
+mydat<-data.frame(x1=c(2,2,3,3,1,1), x2=c(.3, 1, 2.1, 2.2, .1, .2), x3=0.01*c(1,11,4,2,10,6))
+mydat
+
+#The R working directory
+
+getwd()
+setwd("c:/otherdatadir")
+mydat<-read.table("Data/filename.mydat", header=TRUE)
+
+# The read.table()| function 
+
+options(deparse.cutoff=48)
+args(read.table)
+dat<-read.table("Data/testdat1.dat", header=TRUE, skip=5, nrow=2)
+dat
+
+dat<-read.table("Data/testdat2.dat", header=TRUE, na.strings=".", 
+                comment.char=";", dec=",")
+dat
+
+#Reading from more complicated files
+
+cat(1:5+pi,'\n',file="scantest.txt")
+
+vec<-scan("scantest.txt")
+vec
+
+cat(paste("A B C\n",paste(1:5+0.324654, collapse=" "),'\n',"How many roads\n", sep=""),file="readlinestest.txt")
+
+vec<-readLines("readlinestest.txt")
+vec
+strsplit(vec[2]," ")
+as.numeric(strsplit(vec[2]," ")[[1]])
+
+#File connections
+
+f1<-file("readlinestest.txt", open="r")
+scan(f1,what="",nlines=1)
+scan(f1,what=double(),nlines=1)
+readLines(f1)
+close(f1)

5.Reading in Data/Reading in Data.pdf

@@ -0,0 +1,68 @@
+#Writing data to a text file}
+
+mydat<-data.frame(x1=c(2,2,3,3,1,1),
+                  x2=c(0.3,1,2.1,2.2,0.1,0.2),
+                  x3=c(0.01,0.11,0.04,0.02,0.1,0.06))
+mydat
+
+write.table(mydat,file="c:/datadir/filename.dat"),
+row.names=FALSE,col.names=FALSE,sep=", ")
+
+write.table(mydat,file="Data/write.datatest.txt",
+            row.names=FALSE,col.names=FALSE,sep=", ")
+
+#Basic writing functions
+
+set.seed(832)
+cat("Test file for cat\n",round(rnorm(5),3),"\n", 
+    file="cattest.txt")
+
+lin<-c("Count down", paste(rev(1:10), collapse="-"), 
+       "Go")
+writeLines(lin, con="Data/writelinestest.txt")
+
+#A few special ones
+
+sink("Data/sinktest.txt")
+x<-1:5
+y<-1:3
+outer(x,y)
+sink()
+
+set.seed(543)
+x<-1:3
+y<-rpois(10, 4)
+dump(c("x","y"), file="Data/dumptest.txt")
+
+lis<-list(x=1:5, y=3, z=c("a","b","c"))
+dput(lis, file="Data/dputtest.txt")
+dget("Data/dputtest.txt")
+
+#Using file connections}
+
+set.seed(713)
+f2<-file("Data/testout.txt", open="w")
+cat("Header of file\n\n", file=f2)
+mat<-matrix(round(rnorm(12),8), ncol=3)
+write.table(mat, file=f2, row.names=FALSE, col.names=FALSE)
+close(f2)
+
+#Using append=TRUE
+
+set.seed(484)
+
+cat("Header of file\n\n", file="Data/testappend.txt")
+mat<-matrix(round(rnorm(12),8), ncol=3)
+write.table(mat, file="Data/testappend.txt", row.names=FALSE, 
+            col.names=FALSE, append=TRUE)
+
+#Working with binary files: Using save() and load()
+
+x<-rnorm(3)
+lis<-list(y=1:5, z="lalala", fun=function()cat("ha-ha-ha\n"))
+save(x,lis, file="Data/test1.RData")
+
+rm(list=ls())
+load(file="Data/test1.RData")
+ls()
+

6.Writing Data to Text Files/Writing Data to Text Files.R

@@ -0,0 +1,105 @@
+#Reading from SQL databases}
+
+library(RODBC)
+connStr <- paste(
+  "Server=msedxeus.database.windows.net",
+  "Database=DAT209x01",
+  "uid=RLogin",
+  "pwd=P@ssw0rd",
+  "Driver={SQL Server}",
+  sep=";"
+)
+
+if(.Platform$OS.type != "windows"){
+  connStr <- paste(
+    "Server=msedxeus.database.windows.net",
+    "Database=DAT205x01",
+    "uid=PBIlogin",
+    "pwd=P@ssw0rd",
+    "Driver=FreeTDS",
+    "TDS_Version=8.0",
+    "Port=1433",
+    sep=";"
+  )    
+}
+
+conn <- odbcDriverConnect(connStr)
+
+\end{itemize}
+\end{itemize}
+\end{frame}
+
+
+\begin{frame}[fragile]%%\linespread{0.9}
+\frametitle{Connecting to a local SQL Database on your harddisk:}
+\begin{itemize}
+\item Replace server name with the SQL server name on the local machine;
+
+\item With the default SQL installation, this is equal to the {\bf name of the local machine}:
+  
+  
+  \begin{Sinput}
+>connStr <- paste(
+  +    "Server=My_Machine",
+  +    "Database=DAT205x01",
+  +    "uid=PBIlogin",
+  +    "pwd=P@ssw0rd",
+  +    "Driver={SQL Server}",
+  +    sep=";"
+  +    )
+
+connStr <- paste(
+  "Server=msedxeus.database.windows.net",
+  "Database=DAT209x01",
+  "uid=PBIlogin",
+  "pwd=P@ssw0rd",
+  "Driver=FreeTDS",
+  "TDS_Version=8.0",
+  "Port=1433",
+  sep=";"
+)
+
+conn <- odbcDriverConnect(connStr)
+
+#A first query
+
+tab <- sqlTables(conn)
+head(tab)
+
+#Getting a table
+
+mf <- sqlFetch(conn,"bi.manufacturer")
+mf
+
+#Submit real SQL
+
+query <- "
+SELECT Manufacturer
+FROM   bi.manufacturer 
+WHERE  ManufacturerID < 10
+"
+sqlQuery(conn, query)
+
+#Large tables
+
+sqlQuery(conn, "SELECT COUNT(*) FROM bi.salesFact")
+sqlColumns(conn,"bi.salesFact")[c("COLUMN_NAME","TYPE_NAME")]
+sqlQuery(conn, "SELECT TOP 2 * FROM bi.salesFact")
+
+df <- sqlQuery(conn, "SELECT * FROM bi.salesFact WHERE Zip='30116'")
+dim(df)
+
+sapply(df, class)
+
+#SQL summary statistics
+
+df <- sqlQuery(conn,
+               "SELECT    AVG(Revenue), STDEV(Revenue), Zip
+   FROM      bi.salesFact
+   GROUP BY  Zip"
+)
+colnames(df) <- c("AVG(Revenue)", "STDEV(Revenue)", "Zip")
+
+close(conn)
+
+