Suppose you want to load a table named ‘t1’ in a database named ‘db1’ and use a load file called ‘t1.tbl’. The file has the following data and uses the pipe delimiter to delimit the data: 
In the current version of InfiniDB, the file needs to be in the data import directory, which on normal Linux installs is /usr/local/Calpont/data/bulk/data/import. Also the utility currently requires the naming format of
The first step is to execute the colxml utility (found in the Calpont bin directory) which takes the flat file and readies it to be loaded into InfiniDB’s column-oriented structure. The following example creates a job file for the ‘db1’ database and the ‘t1’ table, with the number ‘501’ as the load job for this particular load operation:
Once the job file has been created, you then execute the cpimport utility (found in the Calpont bin directory) to actually load the data into the table:
You can also load multiple tables in one bulk load job. For example, suppose you want to load data into the following tables that comprise a database name ‘tpch2’:
You can use the default pipe (‘|’) field delimiter in their load files, place the load files for all the tables in the import directory, and then execute the prepare for the load job as such for the ‘tpch2’ database:
Then the actual bulk load operation can be run:
