A Few Questions on the subject of bulk data loading.

Hello. A few questions: Our product, Sagent Data Mart Solution, is an end to end data mart product. An important piece of the "equation" is the ongoing population of the data mart (aka, data movement). A critical factor in data movement is performance (especially when having to "fit" into a nightly batch window). Therefore, for the major databases supported we provide some form of batch loading. "Problem Space" The two most important considerations regarding HOW we provide batch loading are: 1. Ability to integration load process into Sagent - the best method is to take advantage of the database vendor client library's batch load API (if available). Unfortunately, not all database vendors provide such an interface (ex, SQL Server and Sybase do but Oracle does not). In Oracle's case we provide command line execution of Oracle's SQL*Loader batch load utility. The two critical questions for the utility are can accept standard input and/or read from a pipe and the other key question is can it read ASCII delimited data? 2. Performance - this is the most important issue. Not much else needs to be said. Questions: - Does the CLI API provide for the batch loading of data? The LOAD SQL statement? Anything else? - If the LOAD SQL statement is the only means of batch loading via CLI can it handle standard input or reading from a pipe? - Which of the different utilities might be best? dbaccess, dbload, hpl (high performance load) - Must any of the different utilities run on the same machine as the Informix server (ie, can the utility run remotely? what about remotely across platforms - NT to UNIX?) - Will the HPL (7.2/UNIX only) ever run on NT? TIA