Sybase IQ uses column based storage vs the typical row based storage of almost all other DW and RDMS procucts.
That said, how IQ stores the data really doesn't affect load times. The problem you have is that you are inserting rows one at a time.
Sybase IQ is a pure DW product not a tweeked RDMS. It is not designed for high volume OLTP type transactions, ie single row insert/update/delete activies.
To get the performance you need there are two fairly easy ways to get the data into IQ quickly.
1. Use the Sybase Bulk Load or Load Table feature. Create a text file from your source table data and a file that descibes the data structure, see Sybase IQ Documentation. Then execute a sql statement to load the data.
2. Sybase IQ is actually composed of two products Sybase SQL Anywhere and IQ. The SQL Anywhere part manages the DW metadata, think masterdb in ASE. The IQ part manages the actual data storage and the DSS query resolution. The SQL Anywhere part is the Catalog Store. You can create tables there and load data into them as you would a regular RDMS with the throughput you would expect. You can also wipe out your IQ database if you mess-up. Query performance isn't very good for data stored in the Catalog Store compared with the performance of queries resolved from the IQ data store. The way around it is to initially load data into the Catalog Store then move the data into the regular IQ data store by doing a 'insert <IQ data store table name> select * from <catalog store table name>' then delete the catalog store table. Or a safer methodology, if you have a SQL Anywhere or ASE database to use, is to load the data, row at a time, into a temporary table in either of them, then by defining the SQL Anywhere or ASE server as a remote server in IQ you can grab the data out of these remote tables using 'insert <IQ data store table name> location <remote server, dbname> select * from <remote temp table name>'. Either method will get the data into IQ quickly.
