- data.frame
    - kids<-c("Jack","Jill")
    - ages<-c(12,10)
    - d<-data.frame(kids,ages,stringsAsFactors=FALSE)
    - d
        - kids ages
        - Jack 12
        - Jill 10
    - d[[1]]    # "Jack" "Jill"
    - d$kids    # same as above
    - d[,1]    # same as above
    - examsquiz<-read.table("exams",header=TRUE)
    - examsquiz[2:5,]
    - examsquiz[2:5,2]
    - class(examsquiz[2:5,2])    # "numeric"
    - class(examsquiz[2:5,2,drop=FALSE])    # "data.frame"
    - x<-c(2,NA,4)
    - mean(x)    # NA
    - mean(x,na.rm=TRUE)    # 3 or use subset()
    - examsquiz[examsquiz$Exam.1>=3.8,]
    - subset(examsquiz,Exam.1>=3.8)    #ERROR: subset(examsquiz,examsquiz$Exam.1>=3.8)
    - d4
        - kids states
        - Jack CA
        - <NA> MA
        - Jillian MA
        - John <NA>
    - complete.cases(d4)    # TRUE FALSE TRUE FALSE, row view
    - d5<-d4[complete.cases(d4),]
    - d5
        - kids states
        - Jack CA
        - Jillian MA
    - rbind(d,list("Laura",19))
        - kids ages
        - Jack 12
        - Jill 10
        - Laura 19
    - eq<-cbind(examsquiz,examsquiz$Exam.2-examsquiz$Exam.1)
    - class(eq)    # "data.frame"
    - examsquiz$ExamDiff<-examsquiz$Exam.2-examsquiz$Exam.1
    - d$one<-1
    - d
        - kids ages one
        - Jack 12 1
        - Jill 10 1
    - d1
        - kids states
        - Jack CA
        - Jill MA
    - d2
        - ages kids
        - 10 Jill
        - 7 Lillian
        - 12 Jack
    - d3
        - ages pals
        - 12 Jack
        - 10 Jill
        - 7 Lillian
    - merge(d1,d3,by.x="kids",by.y="pals")
        - kids stats ages
        - Jack CA 12
        - Jill MA 10
    - d2a<-rbind(d2,list(15,"Jill"))
        - ages kids
        - 12 Jack
        - 10 Jill
        - 7 Lillian
        - 15 Jill
    - merge(d1,d2a)
        - kids states ages
        - Jack CA 12
        - Jill MA 10
        - Jill MA 15
    - dl<-lapply(d,sort)    # "Jack" "Jill" and 10 12
    - as.data.frame(dl)
        - kids ages
        - Jack 10
        - Jill 12

- example 1
    - all2006<-read.csv("2006.csv",header=TRUE,as.is=TRUE)    # as.is=TRUE opposite to stringsAsFactors=TRUE
    - all2006<-all2006[all2006$Wage_Per=="Year",]    # exclude hourly-wagers, one kind of data cleaning
    - all2006<-all2006[all2006$Wage_Offered_From>20000,]    # exclude weird cases, one kind of data cleaning
    - all2006$rat<-all2006$Wage_Offered_From / all2006$Prevailing_Wage_Amount    # create a new column
    makecorp<-function(corpname) {
        t<-all2006[all2006$Employer_Name==corpname,]    # get sub data
        return(t)
    }
    corplist<-c("MICROSOFT","ms","INTEL","intel","GOOGLE","google")
    for (i in 1:(length(corplist)/2)) {
        corp<-corplist[2*i-1]
        newdtf<-paste(corplist[2*i],"2006",sep="")    # concate
        assign(newdtf,makecorp(corp),pos=.GlobalEnv)

- example 2
    - count.fields("DA",sep=",")
    - all(count.fields("DA",sep=",")>=5)    # if TRUE, ok
    - da<-read.csv("DA",header=TRUE,stringsAsFactors=FALSE)
    - for (col in 1:6)
        - print(unique(sort(da[,col])))
    - mrg<-merge(da,db,by.x=1,by.y=1)    # merge by their ids

- example 3
    - aba<-read.csv("abalone.data",header=T)
    - abamf<-aba[aba$Gender!="I",]    # exclude infants from the analysis
    - lftn<-function(clmn) {
        - glm(abamf$Gender ~ clmn, family=binomial)$coef    # return coef to loall
        - # R can auto return without explict return statement
    - }
    - loall<-sapply(abamf[,-1],lftn)
    - class(loall)    # "glm" "lm"