We Want to Hear From You! What Do You Want to See on the Community? Tell us here!
Forum Discussion
Solved
Increase the filetypes considered in full text search
Today's Dropbox web and Desktop interfaces for full search support just certain filetypes (.doc .docx .docm .gdoc .gsheet .jpg .paper .papert .pdf .png .ppt .pptx .pptm .gif .gslide .rtf .tiff .txt .xls .xlsx .xlsm .vtt). There are, though, other files that have the same structure as some of the formally admitted filetypes and there is no technical reason to exclude them from a full text search. This is the case of the files used by the scientific text processor LaTeX (.bib, .tex, bbl...). Many of them are technically identical to the .txt files. In fact the current full text search is able to find the searched text in some of the files with these extensions although it does not find all the files having the searched text in them. Despite this rather arbitrary behaviour, the technical support team indicates that the problem is simply that the filetype is not supported for full search by the system.
Would it possible to formally include every file which is .txt compatible in the full text search?
5 Replies
simonDBX Thank you for you for your quick response. If you are going to consider certain extensions I would suggest to include the following four filetypes corresponding to LaTeX system: .bib, .tex, .bbl, .bak
You could consider also to test which files that are technically identical to .txt and proceed with the full text search within them in the same way as is done with the .txt files. This feature can be considered controversial as long as a more powerful search engine (advanced search option) is NOT provided by Dropbox as it can result in too many non-relevant files (work and temporary files) found.
Hello simonDBX ,
Is there any news on the indexing of text-based file formats other than .txt?
Specific use case we are running into is that .md/.markdown files are not indexed which we use extensively for documentation purposes.
The problem is made worse by the fact that if a file is renamed from .txt to something else, the existing index is no longer updated upon change of the file, causing incorrect search results.
I would argue that simply all text-based file formats as listed on https://help.dropbox.com/view-edit/file-types-that-preview should be indexed. If you can see it on your screen, most users will expect it to be searchable as well.
Best regards, Brian
- simonDBX
Dropbox Product Manager
Hey brian_g, we currently don't have any immediate plans to expand full text search to include the file types you mentioned however, we are always looking at ways to make our search more powerful so this is something we'll continue to keep top of mind as we evaluate what to prioritize next. Thanks again for bringing this to our attention!
Hello simonDBX , thanks for the reply!
Is there a business reason why you currently don't allow the indexing of human-readable text files like markdown? No changes to the indexer are required, I'd think.
Sounds like an easy win to me and nicely fits in with your current introduction of image text recognition, which addresses exactly the same business case.
And anyway the stale indexes and resulting search errors need some solution. We cannot prevent users from changing file types 😕
Best regards, Brian
