All this will be possible in the new version of MemoQ. There will be both cascading filters (i.e. filters can be applied after filters after filters, e.g. Excel sheet containing HTML source with proprietary tags). Then there is the regex filter, which allows you to create tags from regex on the fly. That is, when you encounter unusual tagging, you do not have to reimport the whole document again - just specify the rule and it will be treated as a tag for the time being.
The release candidate is already available, but I am not sure it can be used as a trial version. The final version will follow soon, I suppose...
I already have a license but when I started the project the RC of MemoQ 5.0 was not available yet.
Looking forward to using the filter when I encounter such stuff next time.