How to do many-to-many merge in SPSS?

[ouonomos]

I would like to merge to files based on a many to many two variables,
neither of which are unique.

The first file has two variables:
--the first variable is unique,
--the second variable is shared by groups of the first variable, so
that none of the values for this variable are necessarily unique.

The second file also has the second variable mentioned above for the
first file.
Again, none of these variables are necessarily unique.

My goal is to get a merged file that multiples all of the first
variables in the first file times each case in the second file, where
the second variable in the first file = the variable in the second
file.

Example:

First File
V1 V2
1 A
2 A
3 A
4 B
5 B

Second File
A
A
B

End result should be:
1 A
2 A
3 A
1 A
2 A
3 A
4 B
5 B

Any help you can give me on many-to-many how-to's will be greatly
appreciated. Please note that I have only SPSS, Excel, and Access at
my disposal.

Regards,

Mark

[Raynald Levesque]

>Example:

**** First, create the 2 data file:

DATA LIST LIST /v1(F8.0) v2(A1).
BEGIN DATA

1 A
2 A
3 A
4 B
5 B

END DATA.
SORT CASES BY v2.
SAVE OUTFILE='c:\temp\datafile1.sav'.

NEW FILE.
DATA LIST LIST /v2(A1).
BEGIN DATA
A
A
B
END DATA.
SAVE OUTFILE='c:\temp\datafile2.sav'.

***** Start the job.
GET FILE='c:\temp\datafile2.sav'.
COMPUTE nobreak=1.

AGGREGATE
/OUTFILE=*
/BREAK=v2
/nb = N(nobreak).
MATCH FILES TABLE=*
/FILE='c:\temp\datafile1.sav'
/BY v2.
WEIGHT CASES BY nb.

* For all practical purposes, the above gives you what you need.
* But in case you need the actual cases (not just a weigthed
file giving you the same number of cases), see
"Expand crosstab data into original data file.SPS" in
http://pages.infinit.net/rlevesqu/SampleSyntax.htm#ReadWriteOrCreateData

HTH

Raynald Levesque rlev...@videotron.ca
Visit My SPSS Pages: http://pages.infinit.net/rlevesqu/index.htm