32015-05-01 13:08:47mandy.vogel@googlemail.com22015-05-01 13:05:39mandy.vogel@googlemail.com12015-05-01 13:03:45mandy.vogel@googlemail.com

A common application of loops is to apply a function to each element of a set of values and collect the results in a single structure. In R this is done by the functions: lapply() - works on elements of a list sapply() - same as lapply but simplify results apply() - works on rows or colums of a matrix or a data frame (or more general on arrays) tapply() - works on groups defined by an index

Given the two R object m and l below use * use lapply() to get the class and the length of each element of l (two steps) * apply() to get the maximum of each column in m ]]> ]]>

]]> ]]> ]]> ]]>

we used dir() (with the arguments pattern, recursive, full.path) to get a list of file names we wanted to read in we learned about lapply() which takes a list l and a function f to perform the function f on every element of the list l so now we combine what we learned to read all files at once

]]> ]]>

source the file containing our function read.file() use lapply() to use read.file() on every entry of the list of file names ]]> ]]>

what we get is a list df.list containing the results: every element of the list is a data frame if read.file() read in successfully the respective file so our variable files contains 195 file names ]]> so df.list contains 195 elements ]]> we can check the class of each of these results again with sapply() ]]>

We learned about three basis functions to combine data frame rbind() - combine two data frames row wise cbind() - combine two data frames column wise merge() - combine two data with respect two one or more identifying columns all of them are binary function so you can not put more than two data frame into it using only these function it would be a tedious and boring work to combine 192 data frames

is also a higher order function (functional) Reduce() uses a binary function (like rbind() or merge()) to combine successively the elements of a given list it can be used if you have not only two but many data frames

first we make up 4 artifical data frames (]]> (]]> (]]> (]]> now we use Reduce() in combination with merge() ]]> and what we get is an empty data frame well this isn't exactly what we wanted, so why? it is because the default behavior of merge() is set all=F, so we get only complete lines which is in this case - none so we have to define a wrapper function which only change this argument to all=T set all to TRUE ]]> which is exactly what we want a second example in combination with rbind() ]]> ]]> ]]> which is exactly what we want

the list ml contains three vectors use lapply() to get the class of each of them then use Reduce() and combination with c() to coerce them into one vector. Of which class is the resulting vector? ]]>

use lapply() to get the class of each of them ]]> then use Reduce() and combination with c() to coerce them into one vector. Of which class is the resulting vector? ]]> ]]> ]]>

We used lapply() and our function read.file() to read all files in files and we got back a list df.list containing 192 data frames ]]> ]]>

now use what we learned about Reduce{} and combining data frames using rbind() to combine these 192 data frames.

]]> ]]> ]]>

so it is recommended to build again a function out of this ]]> ]]> by changing the pattern (passed through to dir() we can limit the read in files to specific time or person\tiny ]]> ]]>

table the Subject column again. What is the problem? ]]> subject and time coded in one variable we create two new variables using the str_split() function (stringr package) because str_split() has a list containing a vector as result we have to use it in combination with sapply() then correct some of the person ids ]]> ]]> a alternative is using again regular expressions using the str_replace() function (again stringr package) str_replace() takes three arguments: the string, the pattern to be replaced and the replacement ]]> ]]> ]]> now table the personid column what is left to do? ]]> ]]>

there are some more wrong person ids: RMK - 011, IJ2K - 017, GA3K - 004, Kj6K - 006. Correct them!

]]> ]]> ]]> ]]>

* now read in the file subjectsdemographics.txt using the appropriate command * join the demographics with our data data frame (there is a little problem left - compare the persid and Subject columns) ]]> ]]> ]]> ]]> ]]> ]]> ]]> ]]>

Just run the code and try to understand it. We will cover the ggplot graphics soon. ]]> attachment:graph1.png so there are problems in coding of the test id we remove the letters at the end using str_replace() ]]> ]]> ]]> ]]> attachment:graph2.png And another one. ]]> attachment:graph3.png