Initially we had to move fast from Mainframe (IBM S/390) to Windows Platforrm and to change from ADABAS to ADABAS-D, we had not the time to spend plenty of time with building XML Files, thinking about how to parse them etc. We only had to move the Data quick from ADABAS to ADABAS-D and to take the System-change from EBCDIC to ASCII.
Time went by and there was the need to also be able to move from ADABAS to ORACLE or from ADABAS-D to ORACLE.
As far as I know, the fastest way to bring data into a database is to use its own loader tools.
But if you want to manipulate the data this is another task which I haven't done by now.
I think some problems will also occur, if you have used the special features of ADABAS lwhich are not foundin any other sql databases.
Normalize, normalize, normalize is the mantra. Using VMS, you have several options. Exisiting ADABAS utilities which are fast and fool proof until a smarter fool shows up. But this will not normalize the data. A better solution is to write a Natural program to extract the fields and write the data to an RMS file which then can be loaded using the Oracle loader. Remember that ADABAS permits repeating fields and groups which Oracle will choke on.....