Invalid Tags Created on Import

[obsidianmd]

Invalid Tags Created on Import

[daledesilva]

Oh, apologies for the confusion, my function is intended to maintain tag functionality. It turns #1920 into #_1920
Google Keep allows labels of any characters, so this was to enable those labels to convert to hashtags.

[egauthier]

As you can imagine, this happens a lot! I have a large number of Notes in Evernote with strings like:

• Suite #200
• Invoice #123
• Order #456

[lishid]

Hmm I think tags in the text itself doesn't allow starting with digits because it is quite likely that they are used for numbering things - however for frontmatter this is ok since we are explicitly stating tags.

[claremacrae]

In case it helps, here is the write-up of tag validity in the docs:
https://help.obsidian.md/Editing+and+formatting/Tags#Tag+format

In fact, from a user issue in Tasks, I think that #12.3 is also not treated as a valid tag.

[egauthier]

correct. According to the documentation

Tags must contain at least one non-numerical character. For example, #1984 isn't a valid tag, but [#y1984] is.

Yet after using importer, I have many tags (which I used in Evernote) that are simply numbers. In my case years

[j-palindrome]

I added some code for the Notion importer to escape most tags, which might be useful here:

export function escapeHashtags(body: string) {
	const tagExp = /#[a-z0-9\-]+/gi;

	if (!tagExp.test(body)) return body;
	const lines = body.split('\n');
	for (let i = 0; i < lines.length; i++) {
		const hashtags = lines[i].match(tagExp);
		if (!hashtags) continue;
		let newLine = lines[i];
		for (let hashtag of hashtags) {
			const hashtagInLink = new RegExp(
				`\\[\\[[^\\]]*${hashtag}[^\\]]*\\]\\]|\\[[^\\]]*${hashtag}[^\\]]*\\]\\([^\\)]*\\)|\\[[^\\]]*\\]\\([^\\)]*${hashtag}[^\\)]*\\)|\\\\${hashtag}`
			);

			if (hashtagInLink.test(newLine)) continue;
			newLine = newLine.replace(hashtag, '\\' + hashtag);
		}
		lines[i] = newLine;
	}
	return lines.join('\n');
}

[claremacrae]

Wow, when I see a regex like that, i.e. any non-trivial regex, my first thought is always 'I do hope the project has thorough unit tests, with a wide range of inputs to test the regex really thoroughly...'

[ericaxu]

@egauthier Can you check where the tags end up? Are they in the frontmatter tags property?

[egauthier]

Obsidian: v1.4.1 (installer version 1.2.8)
Importer: v1.0.4

This screenshot is with both hashtag strings in the body and Evernote tags on the Note itself.

Both are added to the global tags. The hashtag strings are not added to the frontmatter, but are added to the Obsidian tag hierarchy. The #123 in the body did not create a tag, presumably because it is not a valid Obsidian tag, but the 1234 Evernote tag did.

[daledesilva]

In my pull request for Google Keep import, I actually created a sanitizing function to deal with this when I noticed it happening. Simply putting an underscore before any tags that start with a number.

The PR is still in review, but this function could be used for sanitisation of other hashtags in the future and evolved to work differently if needed.

[j-palindrome]

I think a backslash is better than an underscore. \# is escaped in Obsidian and doesn’t show up as a tag. Sent with Spark

…

On Aug 1, 2023 at 5:36 AM -0600, Dale de Silva ***@***.***>, wrote: In my pull request for Google Keep import, I actually created a sanitizing function to deal with this when I noticed it happening. Simply putting an underscore before any tags that start with a number. The PR is still in review, but this function could be used for sanitisation of other hashtags in the future and evolved to work differently if needed. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>