Is DevonThink actually not usable for this situation?

I have had DTP for awhile and not used it much. I finally found what I thought would be a great use—searching through large groups of documentation for a software environment that I use. Unfortunately, it seems to be impossible to do something as basic as searching for an XML tag. Is this actually true?

In other words, I want to find in several thousand pages of documentation. Searching for just the word “feature” returns several hundred results. Is there really no way to search for something that includes punctuation? If so, then the using DT for any computer-related field is going to be pretty limited.

Even if I can bypass the index and do a relatively slow search across the data, I would be satisfied (though not as happy as if DT can actually use its more advanced capabilities in the search).

I’m getting the feeling that I may pretty much have to abandon DTP and just start doing searches directly in 20 or 30 individual PDFs. :frowning:

Thanks for any suggestions.

DEVONthink doesn’t search non-alphanumeric characters (for performance reasons related to the Concordance and AI algorithms).

Which means: Non-alphanumeric characters such as < and > will not be found.

Oh well, I had a feeling that would be the answer. I was just hoping there might be a work around to force a non-performant search against the actual data. It would be nice to keep the files in DTP just for the organizational features.

I’m sure I’ll find other decent uses for DTP. This was just the first use case that I had gotten excited for the possibilities of using in this tool. I’ve had a license for several years, but haven’t gotten around to doing much with it. While it seems good for alpha-numeric data, the searching limits may not actually work for most of the data that I end up searching through; primarily software and hardware documentation.

Hi. I have a similar problem. The software does not handle searches of East Asian languages very well, and it often fails to find content that is easily located with other applications such as Spotlight. The effectiveness of the AI also seems to be negatively impacte. Fortunately, there is a workaround: indexing. If you index your data, you can stil take advantage of DT’s many features while using other applications to conduct searches. In your case, it seems to me that something like BBEdit would work well (you can textify the PDFs if necessary). Here are some links:

That’s a useful comment. A third party app such as BBEdit could be used for searches of Index-captured items ‘marked’ by non-alphanumeric characters, or even used to batch convert those markers to something that DEVONthink can ‘see’ such as nonsense strings zzz and xxx to replace, respectively, < snd > – although I would probably add a space after the nonsense string marker…

thanks. in my experience, search applications / functions all have their quirks / strong points / weak points, because there are necessary tradeoffs to be made. bbedit, for example, doesn’t search filenames, as far as i know. this can be a tough hurdle for someone like me who relies a lot on titles to sort and navigate files. but, dt works just fine with titles. in the end, a mix of overlapping apps works pretty well, and because dt has great support for indexing, there is no walled garden or silo effect to worry about.