Trouble with wildcards about dynamodump HOT 3 CLOSED

Hi Ryan, the code currently assumes the 'environment prefix' is separated from the table name by a -. However, in your case, it is split by . thus the reason why it was not working correctly.

I've clarified this behavior in the readme to make it clear in the latest update to master (ffe4d5f, 0.1.13) and have added a --prefixSeparator optional argument so your command should work correctly now:

python dynamodump.py --mode backup --region us-east-1 --srcTable test* --prefixSeparator .

As for your query on test.atlas, unfortunately this would not be supported currently as the text before the separator is being treated as the environment and this change would impact the overall behaviour (as it seems you're using another namespace level). If you have ideas on how to do this, feel free to open a PR.

from dynamodump.

Thanks for looking into this. Why do you have the namespace rules at all? Why not just use traditional globbing like Unix shells use, and treat all non-wildcards equally?

In my case, I have 7 tables named test.atlas.<something> and 1 table named test.scorekeeper.<something>. I need to export all of the atlas tables but need to skip the scorekeeper one. It has millions of records and an export would take days to finish. I wrote a wrapper script that runs dynamodump on each table separately, but I'd prefer to use the wildcard syntax if I can.

from dynamodump.

It's not a namespace rule per se, just how the script assumes the first part of the table name to be the 'environment' as it was written with the idea of backing up/restoring data between different environments, for simplicity sake.

You're right that it probably can be switched to use wildcards as it is done in traditional globbing, but as I've mentioned before, this involves a change in behaviour and I would not be looking to change it for now. :)

from dynamodump.