Clean reading of HTML table with R

Question

I have some data under an HTML format and would like to import it in R. However, it seems that type inference is much less sophisticated than what exists with functions such as read.table. I am using the following workaround, which consists in reading the data table once, write it as a csv and re-import it.

# import raw data as HTML
require(XML)
raw_data <- readHTMLTable(doc = ".\\data\\mydata", 
                          header = T, 
                          as.data.frame = T)
# clean format and import

write.table(x = raw_data, file = ".\\temp.csv", quote = TRUE)
clean_data <- read.table(".\\temp.csv")

removeDots <- function(x)
{
  x <- gsub(pattern = "NULL", replacement = "", x = x, fixed = T)
  gsub(pattern = ".", replacement = "", x = x, fixed = T)
}

colnames(clean_data) <- lapply(X = colnames(clean_data), FUN = removeDots)

Per example, the following data is imported as a table of factors:

<table style="width:100%">
  <tr>
    <td>Jill</td>
    <td>Smith</td> 
    <td>50</td>
  </tr>
  <tr>
    <td>Eve</td>
    <td>Jackson</td> 
    <td>94</td>
  </tr>
</table>

Is there any better option ?