====== Import Data to R ======

These instructions describe how to import the collected data into [[https://www.r-project.org/|R Project for Statistical Computing]]. You essentially have two options here.

1. you can retrieve the data directly from R.

2. you can download a script and a CSV file in SoSci Survey and import them.

The same thing happens in the background: an R script imports a CSV file. However, the first option is much more convenient.

In both cases, the variables are provided with a description using ''comment()'', and the values of selection questions are stored as ''factor'' unless otherwise set. For scale questions, the answer options are labeled using ''attr()''.

<code R>
> comment(ds$SERIAL)
[1] "Personal identifier or participation code (if used)"


> levels(ds$AB01)
[1] "Yes, regularly"                    "From time to time"
[3] "No or only in exceptional cases"   "[NA] not answered" 

> attr(ds$SC01_01, "1")
[1] "does not apply at all"

</code>

===== Import via API =====

You can use either a temporary or a permanent URL for the API download.

  * Temporary URL:\\ **Collected data** -> **Download data** -> tab //GNU R// -> button //Enable API import//.
  * Permanent URL:\\ **Collected data** -> **Data retrieval via API** -> plus symbol top right -> function //Retrieve data as CSV for R// -> Save\\ (you will then find the command in the line //Command// in the overview)

In both cases, SoSci Survey provides a short code that looks like this

<code R>
eval(parse("https://www.soscisurvey.de/rtr/?act=f20H7pQDcp1OCdseFRmCUSmj&vQuality&rScript", encoding="UTF-8"))
</code>

This code retrieves a script directly from the URL, interprets it as an R script and executes it. Copy this command into your R console or into an R script and execute it.

**Tip:** If you encounter problems, you can also call up the script directly in the browser via the specified URL, copy it to R and execute it there.

<code>
https://www.soscisurvey.de/rtr/?act=f20H7pQDcp1OCdseFRmCUSmj&vQuality&rScript
</code>

===== Import via Download =====
Select the "GNU R" tab under **Collected data** -> **Download data** and download both the R script (R file) and the CSV file ([[download#gnu_r|Download data: GNU R]]).



**Note:** If the data and the script were downloaded at different times, it may be that the files are incompatible (e.g. because the list of variables used may have changed).

**Note:** When downloading the survey data, choose the //GNU R// file format instead of //CSV//. This ensures that the formatting in the file will be compatible with the import script.
 
The following steps describe the process when using the RGui user interface for R. In principle, the process is virtually the same if you are using R from the console, or another GUI such as [[http://www.rstudio.com/|RStudio]].

Open and start the script in R:

  * //File// -> //Open script//
  * //Edit// -> //Run all//
  * A dialog should now open that allows you to select the file (this may be behind the window currently open). Select the CSV file that you downloaded and confirm with //Open//.
  
The survey data should be be located in a data frame that you specified when downloading the script (default is "data"). You can check the variables using the R command ''str(data)''.

===== Notes =====

It has already been pointed out above that the R import labels the variables using ''[[https://www.rdocumentation.org/packages/base/versions/3.6.2/topics/comment|comment()]]''. You can use the ''Varlist()'' function from the ''[[https://github.com/BurninLeo/R-helpers|R-Helpers]]'' to display a variable overview for your ''data.frame'', as is available in SPSS, for example.

To ensure that this information is not lost during subselects, the import script adds an ''avector'' class by default, which copies the attributes during the subselect. If the ''avector'' becomes a problem with a ''bind_rows()'', for example, you can remove this class with the following function.

<code R>
remove_avector = function(x) {
    x = data.frame(
      lapply(x, function(xx) {
        class(xx) = setdiff(class(xx), c("avector"))
        return(xx)
      })
    )
    return(x)
}

dc <- remove_avector(ds)
</code>

When downloading via CSV file and with the temporary API link, you can switch the additional function on and off using the option "Add function to retain variable descriptions in subsets".