What I’m trying to do
Hi
i changed to upsidian recently and imported a few thousand of notes. imported images seem to have URL’s that have thousands of characters (see screenshot), i think, that some things function a bit slow because of that. But the main problem is: these long URL’s contain a lot of character-combinations (s.Screenshot) and they lead for example to hundreds of search results, instead of just a few.
so the search results are unusable for me
Things I have tried
i did not find anything that could help me to get correct/reasonable search results. so i am a bit desperate.
I think one the these approaches could help, maybe on exists:
search for terms without markdown-code included (just viewing mode)
exchange these crazy URL’s from import somehow to something shorter (as a batch job, i have thousands)
…anything else of course, that could help
search exclusive of markdown-code, exchange image url as batch job
For a temporary and immediate solution you can try searching using double quotes like "search term".
This will only give you results that match exactly.
But I’d also suggest changing the links for a better/long term solution.
Remember, you don’t need to do it one shot. You can simply handle the links whenever you come across them in your regular use.
You can use the settings to set custom location for your attachments to keep things organised, or even use plugins like custome attachment location to get more fancy
If those URLs are multiple screens long they may be base64-encoded images (Base64 - Wikipedia). This is a way to encode the images as text so they can be put in a text file. There may be a plugin that can convert them to files, or to rename them, or to download them if they’re on the web. Search “Image” in community plugins to see the options. Or you may be able to drag the pictures out to Obsidian’s or the OS’s file browser to save them as files and re-add them, or maybe you could go back to your exporter or importer, change a setting, and re-export/import.
you guessed right, thank you.
indeed, those are base64-encoded images. thank you.
getting rid of them will, it seems, be quite difficult, as i have thousands of them. One plug-in offers encoding so far, but no decoding - and batch processing might be even more of a hurdle.
(wonder, why the importer does not ask, if this option is favored, when migrating)
As I mentioned before, remember that you don’t have to do it all at once.
Just convert the encoded urls to images as you come across them in your regular review/reading/work.
Admittedly I don’t know your exact work method so it might still be difficult, but it’s something to consider.
yes, as long as there is no plugin or another helper, that batch-processes it, there is no real fix possible.
the very best solution for now/this situation would be, to be able to exclude URL’s from a search.
the proposal above to search for “term” does work better than without quotes (hundreds/thousands of wrong search results) but is too limiting, as it does not find expressions like “terms” or “terminator”
i studied the help and many examples, but there seems no approach working.
it would need something like: term -links (i saw some requested it)
i am really a bit desperate, as the sense of a note app is to get proper results
Another approach would be to attempt a new import. Where did the notes originally come from, and how did you get them into Obsidian? Are these Markdown files that were created by the prior app? Did you use a tool that converts another app’s files to Markdown and then move the results into your vault folder? Did you use the Importer plugin to import whatever the prior app exported?
