adding my vote to this request. essential feature, great for there to be an option to be toggled in preferences, my preference would be for the default to ignore accents (i.e., treat cafe and café as equivalent in search). thanks!
Even as someone who currently only speaks English to a high enough degree of fluency to write notes in, I find this very disruptive. A good chunk of my notes are about topics that use terminology borrowed from other languages that in some way use accents on their letters (either they’re written that way to begin with, or they’re borrowings from languages that don’t use latin at all and need extended latin to properly translitarate the term). Stripping them of their accents in my notes just looks and feels weird. If it’s frustrating to me, I cannot imagine how much of a pain this must be for people who write exclusively in accent-heavy languages! I understand what was mentioned about it being difficult to implement, so I guess I get it if it takes a while to implement, but I still feel like I should voice my support.
As an aside, I know Omnisearch fixes the actual searching problem, but does anyone know of a plugin that fixes the note linking problem? That’s what’s been most disruptive for me.
Thank you for even trying! I’m one of those people (well, not exclusively, but Spanish and Catalan are my main ones, and they’re so accent-heavy that even my name has accents) Anyway, feels good to get empathy from someone who isn’t in the exact same situation. You sound like a good person, ngoomie.
Obsidian team: this is useful software and I’m very thankful for your work (you’re also great), but please, please consider moving this issue up in your priority list, you’ll make a lot of lives much easier. ![]()
This is a big downside for me in Obsidian. It has been 5 years, so I don’t think we’ll get this, but it would be useful!
I don’t use Omnisearch because most of the time it breaks obsidian. For now, I just make lots of aliases for the same word.
Can’t this be a plugin though?
There are plugins that do this. Listed above in this thread. For example: Ignore accents/diacritics in link suggestion, quick switcher, find in file and global search - #89 by youben
I meant a plugin to solve this issue, not Omnisearch or Another Quick Switcher. These plugins do not solve this issue as Obsidian is built around search (search, unlinked mentions, linking or embedding files, blocks or headings, and so on). In my case, the linked thinking approach is half baked, since I never have all the results in unlinked mentions (and aliases), search, switcher, etc. For a word with two diacritics, in this moment I have to do four searches instead of one right now, and create all variants of diacritics in aliases. I don’t know why I’m writing all this.
Sorry, I didn’t mean to discount all those other aspects. This is definitely an important FR, and I hope it gets fully implemented someday too. I don’t personally know the full scope of what a plugin could handle.
One of the most needed features for integral and proper search!
The lack of a diacritic-insensitive/-agnostic search is a huge flaw when working with non-Latin scripts and non-English languages. The amount of workaround is huge: either make every search 2 times (one without diacritica, one with – if you remember all diacriticas correctly) or include every term in 2 versions (one with and one without diacritica).
It’s surprising that this key feature request has received no acknowledgment since 2020, given its importance.
It’s surprising from a use case / user perspective but not from a business-managment perspective. Where lies the money? Not in useful features for power users (that other apps already have) but it sell-able feature and recurring revenue (e.g. sync where you do not allow self-hosted on-premise solutions). In the end it will be a question of how many users are annoyed by these little feature-lacks that in the meantime sum up to a large pile of nuisance. At this point the Obsidian team shows and will further show how user-oriented they really are. I rather pay a little money for user-oriented stuff than having all free and the focus goes on payed features that brings the money in.
Thanks for the reminder of the devs’ explanation. It’s obvious that diacritica-agnostic search is not equally easy in every language but I had thought that at least for latin script with 5 vowels and their variants there must be a generic solution that solves the 80% use case. But, of course, one could be wrong in assuming this since it depends on the specific search implementation of Obsidian.
Hi @CawlinTeffid, sorry what did you mean here ?
You said the issue hadn’t been acknowledged since 2020; I was pointing out a more recent acknowledgement.
I believe we’ve given an attempt to properly handle diacritic ignoring search but because of unicode complexities (sometimes the accents are separate characters, sometimes they are not), the search result highlighting would get improper character offsets (meaning it would highlight the wrong part of the sentence).
Such complexities must occur at language level, and thus search can be optimized for individual languages. Some may opt for multiple languages if they deem fit their work flow.
I use IAST and thus my unicode, like others, is limited. If there were an option where I can set my diacritics use such as s = ś, Ś, ṣ, Ṣ, s, S, I would. Better if I could send a PR for my language (e.g., hi) that could be used by other hi speakers.
Devs, please incorporate this feature. It makes using Obsidian so better for so many. Else, we are just using wrong filenames (i.e., without diacritics) for simplicity.
Update after update and this issue remains unresolved. It increasingly seems it may never be addressed.
There are already open-source plugins that accomplish this task, so it’s unclear why implementing similar functionality in the core would be so difficult.
The only note taking software that cannot handle such-a-basic thing, despite being one of the most popular and widely used.
+1 This would be such a quality of life improvement.
As I write most of my notes in French, it’d be great to see this feature implemented.