CODEWITHSUNDEEP
syntaxspss
Timo
Timo

Reputation: 31

Changing value of existing valuelabels via Syntax

I'm currently transforming my survey results for analysis. The survey program I used (limesurvey) automatically generates value labels, but I need to change the value the labels corresponds with. SPSS sadly doesn't do that automatically with the recode command, which only changes the values but doesn't change the value labels that correspond to it (They keep pointing at the old value which doesn't have any data entries anymore).

Basically I want to change the value labels from this: Value labels before change
to this: Value labels after (Changing the labeled Values from A1,A2,A3,A4,A5 into 1,2,3,4,5)

And I don't just want to do it for a single variable, but dozens of variables. While this can actually be accomplished by the search/replace function via the User Interface of SPSS (that's actually how I did it in the example), I would like to do this via Syntax, so I don't have to handcraft every dataset and don't risk screwing up my data with typing mistakes.

Does anyone know how it can be done with Syntax? (Or alternatively some plugin)

Upvotes: 1

Views: 925

Answers (1)

Timo
Timo

Reputation: 31

So I've found a solution to my problem. I did the recodes and the relabeling with Python.

BEGIN PROGRAM PYTHON3.
import spss, spssaux, re

sDict = spssaux.VariableDict() #Get a copy of the Variable Dictionary 

infotext = "Variables processed: " # begin infotext

#Iterate through all variables (in the dictionary - all variables in this case)
for var in sDict:
    #Only Adress variables which are of type String and have value labels.
    if spss.GetVariableType(var.index) > 0 and var.ValueLabels != {} :
        infotext += spss.GetVariableName(var.index) + ', ' #Add processed variable to infotext
        
        #Begin Value Labels and Recode commands for active variable
        commandlab = 'Value Labels ' + spss.GetVariableName(var.index) + ' '
        commandrecode = 'Recode ' + spss.GetVariableName(var.index) + ' '
        #commandalter = 'Alter Type ' +  spss.GetVariableName(var.index) + '(f2).'
        
        #Open entry for each value of valuelabels
        for key,val in var.ValueLabels.items():

            newkey = re.findall(r'\d+', key)[0] #Get first number in key (value of label)
            newkey = re.sub(r'0+(.+)', r'\1', newkey) #Cut leading zeros from number
            
            #Add entry to recode and Value labels command:
            commandlab = commandlab + newkey + ' ' + '"' + val + '"' + ' '
            commandrecode = commandrecode + '("' + key + '"="' + newkey + '") '
        
        #Complete commands for SPSS:
        commandlab = commandlab + '.'   
        commandrecode = commandrecode + '.\n Execute.'
        
        #Execute recode and relabel.
        spss.Submit(commandlab)
        spss.Submit(commandrecode)
        #spss.Submit(commandalter)
    
#Print Infos
print(infotext)
print("Warning: Only Values with value labels have been recoded. Please check your dataset to make sure all entries have been affected.")
end program.

Basically this program is finding all string variables with value labels, gets the numbers from the value/key and recodes and relabels the variables. It can also automatically turn all variables numeric, but I wouldn't recommend doing that without checking all values got converted first. In my case I merged several datasets and one variable had a value without label as a result.

Does anyone know how to get the values (or frequencies) of a variable? If possible, I'd like to catch that problem by doing the recode independent of the value labels.

Upvotes: 2

Related Questions