The performance of the STAR JOIN merge procedure was carefully re-evaluated during development of IBM SPSS Statistics 24. After review of the code, a significant improvement was achieved in the procedure’s performance. The Chart below shows the dramatic difference observed following this optimization in version 24.
Note that the sample datasets used in this test consisted of 100,000 cases and 300 variables each, and matching was performed on two key variables.
Try merging datasets on IBM SPSS Statistics for yourself by downloading a trial here.
Thanks to Rick M for his help on this Blog post.