Search: title influences search results heavily

Is it possible that only the title is being searched?

I have an academic paper indexed. The name is “a.pdf”. I search for the title of the paper but the paper is not in the top 30 search results.
If i now rename the title of the document to the title of the paper, it is the first entry in the search results.

Is this how it is supposed to work?

The title has a higher priority both on iOS and on the Mac.

Ok but not finding a PDF with a title like that (i.e. the entire phrase is in the document) because the title is literally “a.pdf” seems a bit restrictive. on the mac it shows up as #1 even with “a.pdf” as the title

The rating mechanism is a bit different on Mac and iOS. The title is ranked highes also on iOS, then the content, then the comment.

Would it be possible that both work the same? Because for me it is quite irritating, that both apps work different when searching for something. I expected that they are similar, because they are from the same developer and for me the powerful search is quite a strong selling point for DTPO.

Is it possible, that searching in DTPO interprets the input different that DTTG2? If I understood cgrunenbergs answer to a question from me right, DTPO searches for the exact wording, not like spotlight. The exception is the last word, which is interpreted like it has a * at the end until you press enter (if you have that prefix thing enabled). With DTTG2 I noticed that the behavior is different, because here the last word is interpreted as a full word every time, even if I have not pressed enter…

Unfortunately not. DEVONthink To Go 2 doesn’t use the database & search engine of DEVONthink/agent/sphere as it’s not (yet) suitable for iOS & iPhone/iPads, the search engine on the Mac automatically adapts the rankings. E.g. although metadata like the title usually has a higher priority, this doesn’t have to be always the case.

When you say ‘title’ do you mean the title of the paper or the name of the file?

If the first, how would you explain that a paper whose’s filename is a.pdf isn’t found when one searches for the title of the paper?

If the second (that is if the filename is ranked highest, then the content and then the comment) does that mean that the title of the paper isn’t ranked higher than an arbitrary sentence in a footnote?

That’s correct, title is the name of the document. The actual title in a PDF is treated equally to other text. If you extract the text of a PDF, all text looks equal. For a human the paper title is easy to recognize, for an indexing engine it is not necessarily.

I can understand that. Several document managers (e.g. Zotero) are reasonably good in extracting titles from papers, but I think they don’t do so on iOS (or perhaps they do not extract the title but retrieve it in some other way).

One more question: is the title field of a pdf (as seen in the inspector of Preview) indexed? If so how does it rank?

No, the additional metadata of PDFs is not yet indexed, only: document name, content, comment, and tags.

I am looking to buy DTTG and read a reviewer saying that when searching id does not go through content of the documents.

Is that correct? (hoping not…)

the documents.

Please be specific in your inquiry. There is no singular document in DEVONthink To Go.

If I have a PDF stored in DTTG and inside that documents a line says "Here’s looking at you kid.

Would then a search for the word kid show the PDF in the results?

Yes, it would. PDFs also have an in-document search so that you can easier jump to the desired part of the document.

Thanks, eboehnisch:)

Then I guess that is a buy :partying_face: