Homework3.Rmd

---
title: "Homework3"
author: "Melissa Greenaway"
date: "February 5, 2017"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
setwd("R:/Project/EPAR/Working Files/RA Working Folders/Melissa G/R and Python")

```

## R Markdown

If you have this function:

```{r }
riseToPower=function(base,exponent=2){
  #####
  result=1
  if (exponent > 0){
    for (time in 1:exponent){ #number of times to repeat process
      result=result*base
    }
  }
  #####
  return(result)
}
```

We can create another function to deal with values that are negative:

```{r }
riseToPowerPlus=function(base,exponent=2){
  #####
  result=1
  if (exponent > 0){
    for (time in 1:exponent){ #number of times to repeat process
      result=result*base
    }
  } 
  if (exponent < 0){
    for (time in 1:(-exponent)){
      result=result*base
      result=1/result
    }
  }
  #####
  return(result)
}
riseToPowerPlus(4,-1)
```

This function lets us use one input to retrieve several output elements:
```{r}
factors=function(number){
  vectorOfAnswers=c() # empty vector that will collect output
  for (i in 1:number){
    #if the remainder of 'number'/'i' equals zero...
    if ((number %% i) == 0){ # %% gives remainder
      # ...add 'i' to the vector of factors!
      vectorOfAnswers=c(vectorOfAnswers,i)
    }
  }
  return (vectorOfAnswers) # returning  the vector
}
factors(20) #testing
```

But there's a way we can modify it to reduce the number of iterations in the loop -
with large numbers we may have to iterate many times.
```{r}
factors=function(number){
  vectorOfAnswers=c(number) # empty vector that will collect output
  for (i in 1:(number/2)){
    #if the remainder of 'number'/'i' equals zero...
    if ((number %% i) == 0){ 
      # ...add 'i' to the vector of factors!
      vectorOfAnswers=c(i, vectorOfAnswers)
    }
  }
  return (vectorOfAnswers) # returning  the vector
}

factors(20)
```

We can also avoid negative numbers or zeros as input by modifying the code:
```{r}
factors1=function(number){
  vectorOfAnswers1=c() # empty vector that will collect output
  if (number>0){
    for (i in 1:number){ #change this
      #if the remainder of 'number'/'i' equals zero...
      if ((number %% i) == 0){ # %% gives remainder
       # ...add 'i' to the vector of factors!
       vectorOfAnswers1=c(vectorOfAnswers1,i)
    }
  }
  }
  if (number<=0){
    print ("invalid number")
    return(NA)
  }
  return (vectorOfAnswers1) # returning  the vector
}
factors1(20)
factors1(-20)
```

We can also change the code to only accept positive integer values, and make sure an
error message results for an invalid number.
```{r}
factors2=function(number){
  if (number<=0){
    print("invalid number was input")
    return(NA)
  }
  if (number>0){
    vectorOfAnswers=c() # empty vector that will collect output
    if (is.integer(number)){
      for (i in 1:(number/2)){ #change this
        #if the remainder of 'number'/'i' equals zero...
        if ((number %% i) == 0){ # %% gives remainder
        # ...add 'i' to the vector of factors!
        vectorOfAnswers=c(i,vectorOfAnswers)
      }
    }
    }else if (!is.integer(number)){
      print ("invalid number was input")
      return(NA)
  }
  }
  return (vectorOfAnswers) # returning  the vector
}
factors2(1.2)
factors2(-18)
```

We can also use vectors as inputs for functions, and have data frames as outputs. The function
below uses two vectors as input, and returns a data frame with additional columns.
```{r}
operations=function(aVector, avector2){
  #vector of sum
  sumvector=c()
  sumvector=c(sumvector, aVector+avector2)
  #vector of difference
  diffvector=c()
  diffvector=c(diffvector, aVector-avector2)
  #multiple vector
  multvector=c()
  multvector=c(multvector, aVector*avector2)
  #div vector
  divvector=c()
  divvector=c(divvector, aVector/avector2)
  answerasdf=data.frame(vector1=aVector, vector2=avector2, sum=sumvector, diff=diffvector, multiple=multvector, division=divvector)
  answerasdf
}
vectorx=c(10,20,30,40,50)
vectory=c(20,40,60,80,100)
operations(vectorx, vectory)
```

Finally, if we have unknown functions we can explore how they can be applied to different parts of a data frame, including a cell, column, and the data frame as a whole.
```{r}
# The function:
strangerF=function(x){
  answer=0
  change=T
  for (value in x){
    if (change){
      answer=answer+value
    }else{
      answer=answer-value
    }
    change=!change 
  }
  return (answer)
}
```

Exploration:
```{r}
#import data
#install.packages("readx1")
#library(readxl) # you only need to do this once per session
folderName='Data'
fileName='dataforstrangeF.csv'
fileExcel=file.path(folderName,fileName)
strangedata=read.csv(fileExcel,1) # table '1'
strangedata
#browse data
str(strangedata)

#strangerF(strangedata) # cannot apply directly to data frame
#strangerF(strangedata$vals1)
strangerF(strangedata$vals2) # this works! Applying to a single column of integers
strangerF(strangedata$vals3) # this works! Applying to a single column of integers

#Map(strangerF,strangedata$vals1) # doesn't work
Map(strangerF,strangedata$vals2) #works for column of integers, returns list
#mapply(strangerF,strangedata) # doesn't work
#mapply(strangerF,strangedata$vals1) # doesn't work
mapply(strangerF,strangedata$vals3) #works for column of integers, returns vector

integer=function(x){ # I created this function to see if a filter for data type will work
  if (is.integer(x)){
    return(TRUE)
  }else{
    return(FALSE)}
}
Filter(integer,strangedata) #filtering for datatype works

#apply(strangedata,1,strangerF) # 1 for row, doesn't work
#apply(strangedata,2,strangerF) # 2 for column, doesn't work
#lapply(strangedata,strangerF) # doesn't work
```