Wildcard at the beginning and middle of words

I found the search in DTTG not supporting the wildcard at the beginning and middle of the words. I’m a user from China, I rely on this function when I try to search asian words in devonthink pro. When will DTTG 2 support it? Is there any plan already there?

We are using the SQLite FTS5 backend which, unfortunately, only supports wildcards at the end of words. Sorry. But it’s on our list to check if there are other options to make this work.

I now could still search asian language by wildcard in the front/middle of the words in Devonthink pro. Will DT pro use SQLite FTS5 in the future? If yes, I think I have to transfer to other application in the end.

BTW, why haven’t DT support Asian language for such a long time? Evernote and many other notes apps support asian language perfectly. I think it may be not so difficult.

DEVONthink for Mac will not switch to FTS5 because it is built on its own AI-based search engine. Due to the memory and performance limitations that still exist on iOS (and memory limitations are an integral part of iOS itself) its also unlikely that we’ll be able to port the AI engine to iOS.

Asian languages support is, in fact, complicated as you need a commercial tokenizer that splits character sequences into words. Currently, macOS and iOS don’t offer them so we need a third-party solution. But, it’s on our list, for both Mac and iOS.

How much ram would your AI database need?

stackoverflow.com/questions/2107 … os-app-use

You might be able to eat 1 Gb of ram on newer devices and up to 2 on the 7+. I am really curious how DEVONthink AI database would perform on the newer iOS hardware.

If a second app is active on the device and you use 1 GB you’ll see the door from the outside. You can do it for a few seconds but in general consuming more than a few hundred MB, better less, is not advisable. That’s the iOS world we’re afraid.