Hi,
With data sets in 8 million & 16 Million records counts, I would be very carefule not to run out of memory within your SAS session, and your machine!
Hash Object is memory storage, and has a finite limit. Depending how wide your record in each data set (Keys, Data), you could quickly saturate your memory!!
There are alternatives to Hash Object to handle merging/joining large data sets like yours, but they come with their own price tag (Memory/CPU/Disk/Code length and complexity).
Choosing the Right Technique to Merge Large Data Sets Efficiently
In the past, I have used a single data step with mixture of several hash objects and (Set Key=) statements to handle Fact Table to Multiple Dimension tables joins in a single data pass/read.
In short, sometimes, the best solution is a mixture of techniques rather than single technique.
Good luck,
Ahmed
... View more
