I have a customer who is currently running a DNN Repository based system. This was originally intended to be quite small, but has now expanded to over 4000 files and the repository module is inadequate to manage this. So we are looking at putting in DMX and so far it seems ideal - addresses all the shortfalls they are currently experiencing.

My question is (and this may not be the correct forum but none of the others seemed to match ): has anyone any experience with, or suggestions for, performing a migration from the DNN repository?  Ideally I'd like to be able to run this as a scripted process which reads out of the existing site database and populates the DMX tables directly rather than have to download and upload about 60 Gb of data.

Any suggestions gratefully received.

TIA,

Graham

Hi Graham,

There is an import feature in DMX. If you're on a folder it should be visible on the folder menu. It imports from the drive of the web server (or a UNC path). Wouldn't this suffice?

Peter

Peter,

Not really for at least a couple of reasons:

DNN repository obfuscates the file names on disk as a security measure similarly to DMX  - it puts a guid in, so a file whose 'real' name is say 'spreadsheet.xls' is stored as 'spreadsheet.f4b736d0-5814-4868-b545-26a1c9bbfe28.xls'. DMX of course this shows the file as 'spreadsheet.f4b736d0-5814-4868-b545-26a1c9bbfe28' after import which is not very user friendly (to put it mildly ).

We also have categories and attributes associated with many of the files which it would be nice to preserve rather than have to go through a re-entry process.

So unless there is something magic you can suggest to at least strip out the GUIDs I don't think a simple import will suffice. I am envisaging writing a script that exports data from the repository tables to a spreadsheet which will allow some manipulation, then re-importing to the DMX tables, hopefullly using existing stored procedures with as little direct table manipulation as possible to avoid the possibility of corruption. In parallel I can VBscript a filestore copy which will produce 'nice' file names. I have successfully adopted a similar approach with migrating data from an obsolete store to our current preferred product. I haven't looking very deeply into the DMX tables structure yet so I don't yet know whether this will be easy, hard or (next to) impossible. Do you have a schema diagram or similar you can let me have as that will be of some help?

I was sort of hoping that maybe someone had done something similar already and had some scripts they would be able / willing to share. Yes, I am an optimist! As before any advice and guidance would be gratefully accepted.

Regards,
Graham

Hi Graham,

I haven't done anything of the sort, no. But here are my thoughts on this:
- You can directly manipulate SQL data to your advantage maybe. The OriginalFilename column in DMX_Entries and the TITLE texts in DMX_ShortTexts will need to have the GUID taken out. With some SQL magic that might be possible.
- Importing categories and attributes is something you'll also need to check out if it can be done in SQL. If you have a table where the original filename is and the correct data (so do this before the step above maybe) then you might be able to pull this off. Categories themselves are in DMX_Categories and the binding to files is in DMX_EntryCategories. Similarly with Attributes.

Good luck,

Peter