www.infinidb.org

I wanted to offer another InfiniDB load rate metric using the SSB lineorder fact table.  In this case we are using a scale factor of 10,000 which translates to 60 Billion rows.  As a point of reference, the recent Percona benchmark was at a scale factor of 1000 (6 billion rows) http://www.mysqlperformanceblog.com/2010/01/07/star-schema-bechmark-infobright-infinidb-and-luciddb/ .  The load rate per hour varied only slightly across the entire run, averaging about 478 million rows per hour.  As always, your actual load rate will vary based on your disk, table, and column definitions, but you should expect consistent load times across very wide cardinality ranges. 

The table is the Lineorder table as defined here:

Some disclaimers about the SSB at 10,000.  Use of the benchmark at this scale  appears to be bleeding edge:
NOTE: Data generation for scale factors >  1000 GB is still in development,
        and is not yet supported.
Your resulting data set MAY NOT BE COMPLIANT!
In addition, I munged one of 2560 files prior to import and so I am missing about 24 million records vs. the expected.   

mysql> desc ssb_10k.lineorder;
+------------------+---------------+------+-----+---------+-------+
| Field            | Type          | Null | Key | Default | Extra |
+------------------+---------------+------+-----+---------+-------+
| lo_orderkey      | bigint(20)    | YES  |     | NULL    |       |
| lo_linenumber    | int(11)       | YES  |     | NULL    |       |
| lo_custkey       | int(11)       | YES  |     | NULL    |       |
| lo_partkey       | int(11)       | YES  |     | NULL    |       |
| lo_suppkey       | int(11)       | YES  |     | NULL    |       |
| lo_orderdate     | int(11)       | YES  |     | NULL    |       |
| lo_orderpriority | char(15)      | YES  |     | NULL    |       |
| lo_shippriority  | char(1)       | YES  |     | NULL    |       |
| lo_quantity      | decimal(12,2) | YES  |     | NULL    |       |
| lo_extendedprice | decimal(12,2) | YES  |     | NULL    |       |
| lo_ordtotalprice | decimal(12,2) | YES  |     | NULL    |       |
| lo_discount      | decimal(12,2) | YES  |     | NULL    |       |
| lo_revenue       | decimal(12,2) | YES  |     | NULL    |       |
| lo_supplycost    | decimal(12,2) | YES  |     | NULL    |       |
| lo_tax           | decimal(12,2) | YES  |     | NULL    |       |
| lo_commitdate    | int(11)       | YES  |     | NULL    |       |
| lo_shipmode      | char(10)      | YES  |     | NULL    |       |
+------------------+---------------+------+-----+---------+-------+

Example syntax to import:
     /usr/local/Calpont/bin/colxml ssb_10k -t lineorder -j 10000
     /usr/local/Calpont/bin/cpimport -j 10000

Actual row count: 59,977,404,781

Example data: 

|1|1|73799965|1551894|41364203|19960102|5-LOW|0|17.00|3307894.00|18660018.00|
4.00|3175578.00|116749.00|2.00|19960212|TRUCK|

A stable and predictable load rate can be very important when dealing with larger and larger data sets. Let us help you put your data to work!

Thanks - Jim Tommaney,  Chief Product Architect