[chrisplough]

soujanya,

Hello! I've found that CSVs are preferable when working with bulk data loads, because they are easier to work with, are much smaller than XML files (which can get HUGE) and because you can also use tools like SQL*Loader to accelerate the process. You can also include several CSV files into a single ZIP file and control the loading order via csvutil.ctl file - making the loading process even easier.

During most implementations, CSV data imports are used extensively to load the initial data sets and import data over from legacy systems.

I've found little use for importing data via XML (or DB.XML). Instead, I use the XML and web services interfaces when integrating OTM with 3rd party OMS, WMS and other systems.

Hope this helps!

--Chris