Unfortunately I do not have before and after data size numbers fro the row compression on our production OLTP database. But I did use (as documented in the script) sp_estimate_data_compression_savings stored procedure to identity potential tables elgible for row compression (you can also use it for page compression). Examples are indicated below (table names changed to protect the innocent). Of course, the amount of compression is very dependent on the table architecture and also the data within it. However, I was not looking so much for saving disk space as I was for decreasing disk IO (and the attendant performance gains associated with less IO). I guess I'll have to do a future article on potential performance gains from data compression; thanks for pointing this out.

Table IndexID PartitionNumber CurrentSize(kb) CompressedSize(kb) %Compression

AuditLog1 1 68432 47240 31.0

Barcode 1 1 30112 20576 31.7

Barcode 1 2 656592 453528 30.9

Barcode 1 3 707480 488936 30.9

Barcode 1 4 683672 473000 30.8

Barcode 1 5 693344 481232 30.6

Barcode 1 6 755952 525368 30.5

Barcode 1 7 848928 589064 30.6

Barcode 1 8 231976 160808 30.7

Batch 1 1 65184 43872 32.7

USPSPackage 1 1 2705768 2059016 23.9