[macxcited]

Yes, the complete header row of headings must be included, exactly as they are in the export file.

I should have said in my notes above, that when I went back to the original export file and included all the ten columns, resulting in most of them being blank, I never altered the header row from the export file at all.

So, to clarify/confirm, my method is –
1. Add a couple of machines by using WGMgr
2. Export those machines to get the export-import ‘template’ data file
3. Use a ‘unix friendly’ text editor (to not mess with the returns at the end of lines, I used TextWrangler) to edit the data file to include the complement of machines to be imported
4. Import your data file. Expectation is that you should _not_ get presented with a window requesting any field mapping be done; the import should just show a progress bar and complete successfully. If I do get the request for field mapping, I cancel the import, and go back to the data file and check the structure of the rows against the original export file; turn on all invisibles and show spaces etc. in TextWrangler.

Good luck. Cheers, Peter B.

[macxcited]

Managed to get this done just yesterday.
Comment above was helpful. The Admin guides from Apple do have some interesting tips to bear in mind along with detailed info about the classes and OD structure. But, not being a trained programmer or experienced scripter, I found the lack of a “recipe” a bit disappointing. What I was looking for was a ‘how to’ type doco from the perspective of using the GUI available in WkGrpMgr and its Import menu command. I hope the sketchy notes below might suffice.
Note that I was using Server 10.5.6 OD and this was with no AD integration, but for the task of importing machines into the OD I think that should not matter.
Firstly, not knowing the complete structure required for the import text file, in WkGrpMgr I added a couple of bogus machines by hand to the master computer list, entering just the “Name”, “Short Name” and “Ethernet ID”. Then I Exported the data for those machines to see what the export file looked like.
I then edited the export file to include only the fields that I wanted to – RecordName, RealName, ENetAddress – for the list of new machines to be added. I found that in doing this I always got a ‘field mapping’ dialog asking me which data elements from my import file I wanted to map to which fields in the OD. Whenever this occurred, the data would always end up in the wrong OD fields no matter how I set up the mapping. I figured that the Import must require exactly the same field list (columns) that the Export spits out. So I went back to the original Export data file and added the new machine records to it without removing any of the columns that would be empty. Note that I left the column for the GUID empty and only populated the three fields I had data for, but left the full complement of ten data columns intact and in their original order.
Voila! It worked!!!
I also noticed, the extra screen requesting the Field mapping exercise never appeared. OD was obviously happy with the setup of the file and just imported it, no questions asked!
Hope this helps.

[Author]

Posts