I noticed that in the Data Loader mapping file (.sdl) names of fields of target objects must have exact the same spelling as on target Org (mean case-sensitive). Data loader does not show any error or warning but values for fields with names in different case (UPPERCASE for instance) ignores (WHY?) by Data Loader.
The other curious thing is that when you extract data from the org you get csv file where field names are always in UPPERCASE (WHY?).
So, imagine that I need to extract data from one org, automatically generate mapping based on header of CSV-file (here is a problem) and load data to another org with the same data structure using Data Loader CLI.
Do anybody know how to force Data Loader to extract CVS with exact field names (no UPPERCASE) in header? Is it possible to force Data Loader CLI load field values even if names of fields in mapping file specified in different case?