---
title:
author:
date:
output:

pdf_document: default
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
```

```{r, include=FALSE}
library(XML)
ubase = ""

urls = paste(ubase, "results/", 2011:2012, "/",
         2011:2012, "cucb10m-f.htm", sep = "")

extractResTable =
# Retrieve data from web site, find preformatted text,
# return as a character vector from each url.
function(url)
{
   doc = htmlParse(url)
   preNode = getNodeSet(doc, "//pre")
   txt = xmlValue(preNode[[1]])
   els = strsplit(txt, "\r\n")[[1]]

   return(els)
}

womenTables = lapply(urls, extractResTable)
womenTables[[1]][1:10]
womenTables[[2]][1:10]
sapply(womenTables, length)

years = 2011:2012
names(womenTables) = years
names(womenTables)

els2011 = womenTables[[1]]
els2012 = womenTables[[2]]
head(els2011)
head(els2012)
```

```{r, include=FALSE}
eqIndex = grep("^===", els2012)
eqIndex
first3 = substr(els2012, 1, 3)
which(first3 == "===")

spacerRow = els2012[eqIndex]
headerRow = els2012[eqIndex - 1]
body = els2012[ -(1:eqIndex) ]

headerRow = tolower(headerRow)
headerRow

ageStart = regexpr("ag", headerRow)
ageStart

age = substr(body, start = ageStart, stop = ageStart + 1)
head(age)

```

```{r, include=FALSE}
summary(as.numeric(age))

blankLocs = gregexpr(" ", spacerRow)
blankLocs

searchLocs = c(0, blankLocs[[1]])

Values = mapply(substr, list(body),
            start = searchLocs[ -length(searchLocs)] + 1,
            stop = searchLocs[ -1 ] - 1)

findColLocs = function(spacerRow) {

spaceLocs = gregexpr(" ", spacerRow)[[1]]
rowLength = nchar(spacerRow)

if (substring(spacerRow, rowLength, rowLength) != " ")
   return( c(0, spaceLocs, rowLength + 1))
else return(c(0, spaceLocs))
}

selectCols =
function(colNames, headerRow, searchLocs)
{
   sapply(colNames,
            function(name, headerRow, searchLocs)
            {
               startPos = regexpr(name, headerRow)[[1]]
               if (startPos == -1)
                  return( c(NA, NA) )

               index = sum(startPos >= searchLocs)
               c(searchLocs[index] + 1, searchLocs[index + 1] - 1)
            },
            headerRow = headerRow, searchLocs = searchLocs )
}

searchLocs = findColLocs(spacerRow)
searchLocs
ages = selectCols("ag", headerRow, searchLocs)
ages = mapply(substr, list(body),
         start = ages[1,], stop = ages[2, ])

summary(as.numeric(ages))
```

```{r, include=FALSE}
shortColNames = c("name", "home", "ag", "gun", "net", "time")

locCols = selectCols(shortColNames, headerRow, searchLocs)
locCols

Values = mapply(substr, list(body), start = locCols[1, ],
            stop = locCols[2, ])

class(Values)

colnames(Values) = shortColNames
head(Values)

tail(Values)[ , 1:3]

extractVariables =
function(file, varNames =c("name", "home", "ag", "gun",
                           "net", "time"))
{
   # Find the index of the row with =s
   eqIndex = grep("^===", file)
   # Extract the two key rows and the data
   spacerRow = file[eqIndex]
   headerRow = tolower(file[ eqIndex - 1 ])
   body = file[ -(1 : eqIndex) ]

   # Obtain the starting and ending positions of variables
   searchLocs = findColLocs(spacerRow)
   locCols = selectCols(varNames, headerRow, searchLocs)

   Values = mapply(substr, list(body), start = locCols[1, ],
                     stop = locCols[2, ])
   colnames(Values) = varNames

   invisible(Values)
}
data=sapply(womenTables, length)
data

womenResMat = lapply(womenTables, extractVariables)
length(womenResMat)

sapply(womenResMat, nrow)

age = sapply(womenResMat,
         function(x) as.numeric(x[ , 'ag']))

str(age)
```

1) Box Plot of Age by Year for Female Runners

```{r, echo=FALSE}
boxplot(age, ylab = "Age", xlab = "Year")

```

```{r, include=FALSE}
#Start to cope with the times
charTime = womenResMat[['2012']][, 'time']
head(charTime, 5)
tail(charTime, 5)

#Separate the time into list of list
timePieces = strsplit(charTime, ":")
timePieces[[1]]

#2 hours 30 minutes 59 sec
tail(timePieces, 1)

timePieces = sapply(timePieces, as.numeric)
head(timePieces)
runTime = sapply(timePieces,
            function(x) {
                  if (length(x) == 2) x[1] + x[2]/60
                  else 60*x[1] + x[2] + x[3]/60
            })
summary(runTime)

convertTime = function(time) {
timePieces = strsplit(time, ":")
timePieces = sapply(timePieces, as.numeric)
sapply(timePieces, function(x) {
   if (length(x) == 2) x[1] + x[2]/60
   else 60*x[1] + x[2] + x[3]/60
})
}

createDF =
function(Res, year, sex)
{
   # Determine which time to use
   useTime = if( !is.na(Res[1, 'net']) )
         Res[ , 'net']
   else if( !is.na(Res[1, 'gun']) )
         Res[ , 'gun']
   else
         Res[ , 'time']

   runTime = convertTime(useTime)

   Results = data.frame(year = rep(year, nrow(Res)),
                        sex = rep(sex, nrow(Res)),
                        name = Res[ , 'name'],
                        home = Res[ , 'home'],
                        age = as.numeric(Res[, 'ag']),
                        runTime = runTime,

Use the Race Results 1999-2012 from the Cherry Blossom data base to...