- There is one serious error:
The "lastpage" field doesn't get exported in the .csv file. In the R syntax file, the column numbering doesn't keep proper track of this. So, in the second line of the R syntax file, the columns get numbered sequentially:
names(data)=paste("V",1:dim(data)[2],sep="")
But the remainder of the script works with the field names as supplied by the spss_fieldmap() function. This means the data and the syntax file gets out of step (notice the missing V15):
data[,which(names(data)=="V14")]=as.character(data[,which(names(data)=="V14")])
data[,which(names(data)=="V16")]=as.character(data[,which(names(data)=="V16")])
The same error will recur for each hidden variable in the survey.
The fix for this is to have an indexing variable, e.g. $i that gets incremented when the foreach loop outputs something, and to use this indexing variable to refer to the columns.
- A second big problem is that the second to last line in the syntax file is written as one long string (v.names=(c....))
This possibly works fine when calling source(), but when loading into an editor, the line gets truncated.
The solution to this is to add a \n line break character between each column name.
- In addition, there are some encoding problems, meaning that a ' (single apostrophe) and - (minus) gets incorrectly encoded and escaped:
' gets encoded as ’
The solution to this is to use a combination of addslashes() and iconv().
- Finally, some relatively minor R issues. Although the = assignment operator works, it is actually better practise to use the <- assignment operator.
| 
