SPSS TUTORIALS
Running syntax over several SPSS data files in one go is fairly easy. If we use SPSS with Python we don't even have to type in the file names. The Python os (for operating system) module will do it for us.
Try it for yourself by downloading spssfiles.zip. Unzip these files into d:\spssfiles as shown below and you're good to go.
Find All Files and Folders in Root Directory
The syntax below creates a Python list of files and folders in rDir, our root directory. Prefixing it with an r as in r'D:\spssfiles' ensures that the backslash doesn't do anything weird.
begin program.
import os
rDir = r'D:\spssfiles'
print os.listdir(rDir)
end program.
Result
Filter Out All .Sav Files
As we see, os.listdir() creates a list of all files and folders in rDir but we only want SPSS data files. For filtering them out, we first create and empty list with savs = []. Next, we'll add each file to this list if it endswith(".sav").
begin program.
import os
rDir = r'D:\spssfiles'
savs = []
for fil in os.listdir(rDir):
if fil.endswith(".sav"):
savs.append(fil)
print savs
end program.
Using Full Paths for SPSS Files
For doing anything whatsoever with our data files, we probably want to open them. For doing so, SPSS needs to know in which folder they are located. We could simply set a default directory in SPSS with CD as in
CD "d:\spssfiles".
However, having Python create full paths to our files with os.path.join() is a more fool proof approach for this.
begin program.
import os
rDir = r'D:\spssfiles'
savs = []
for fil in os.listdir(rDir):
if fil.endswith(".sav"):
savs.append(os.path.join(rDir,fil))
for sav in savs:
print sav
end program.
Result
Have SPSS Open Each Data File
Generally, we open a data file in SPSS with something like
GET FILE "d:\spssfiles\mydata.sav".
If we replace the file name with each of the paths in our Python list, we'll open each data file, one by one. We could then add some syntax we'd like to run on each file. Finally, we could save our edits with
SAVE OUTFILE "...".
and that'll batch process multiple files. In this example, however, we'll simply look up which variables each file contains with spssaux.GetVariableNamesList().
begin program.
import os,spss,spssaux
rDir = r'D:\spssfiles'
savs = []
for fil in os.listdir(rDir):
if fil.endswith(".sav"):
savs.append(os.path.join(rDir,fil))
for sav in savs:
spss.Submit("GET FILE '%s'."%sav)
print sav,spssaux.GetVariableNamesList()
end program.
Result
Inspect which Files Contain “Salary”
Now suppose we'd like to know which of our files contain some variable “salary”. We'll simply check if it's present in our variable names list and -if so- print back the name of the data file.
begin program.
import os,spss,spssaux
rDir = r'D:\spssfiles'
findVar = 'salary'
savs = []
for fil in os.listdir(rDir):
if fil.endswith(".sav"):
savs.append(os.path.join(rDir,fil))
for sav in savs:
spss.Submit("get file '%s'."%sav)
if findVar in spssaux.GetVariableNamesList():
print sav
end program.
Result
Circumvent Python’s Case Sensitivity
There's one more point I'd like to cover: since we search for “salary”, Python won't detect “Salary” or “SALARY” because it's fully case sensitive. I you don't like that, the simple solution is to convert all variable names for all files to lower()case.
A basic way to change all items in a Python list is
[i... for i in list]
where i... is a modified version of i, in our case i.lower(). This technique is known as a Python list comprehension and the syntax below uses it to lowercase all variable names (line 13).
begin program.
import os,spss,spssaux
rDir = r'D:\spssfiles'
findVar = 'salary'
savs = []
for fil in os.listdir(rDir):
if fil.endswith(".sav"):
savs.append(os.path.join(rDir,fil))
for sav in savs:
spss.Submit("get file '%s'."%sav)
if findVar.lower() in [varNam.lower() for varNam in spssaux.GetVariableNamesList()]:
print sav
end program.
Note: since I usually avoid all uppercasing in SPSS variable names, the result is identical to our case sensitive search.
Thanks for reading.
THIS TUTORIAL HAS 4 COMMENTS:
By Amy on January 25th, 2018
Very helpful page Reuben - can this be done for .sps files? i.e. could we run syntax that would look through the syntax and change x value for y value?
By Ruben Geert van den Berg on January 25th, 2018
Hi Amy!
Well, sort of: .sps files are just plain text files and Python can read/write them but not modify them. However, you could first read the contents, change the contents, delete the file and save the modified contents under the old file name. Or save the modified file contents with a different file name.
Hope that helps!
By Amy on January 26th, 2018
Thanks Reuben - there are potentially hundreds of them, so am looking for a way to automate it, but maybe it is not possible?
By Ruben Geert van den Berg on January 26th, 2018
Hi Amy, it's perfectly possible. It just takes a bit of effort to set up properly.
I thought I'd actually built an SPSS custom dialog for it that can be used from the SPSS menu but I can't seem to find it anymore. I could rebuild it for you but I'd have to charge a couple of hours for that.