Issue with "|" pipe character in Page names

[GitMurf]

From @phhq "EA" in Slack:

I had a few minutes to test it with a csv of six rows of "purchases". I found it failed on two of the six rows the first time, but the other four were fine. I noticed both the page names of the failed rows contained a "|" character. When I removed the pipe character, the conversion worked. I'm not sure if there is an issue with "|" in a page name, or running through power script, etc. -- have a look.

[GitMurf]

@phhq here is the answer. In Windows you cannot have a "|" pipe in the file name (see screenshot below). The other characters not allowed are: \ / : * ? " < > |

Now my question is, what is an acceptable solution for you? I would prefer to just replace any/all of those characters listed above with the same replacement. The question is, what character would you vote for as replacement for each of the illegal characters? Please select all that you would be fine with or just One if you have a strong opinion about it.

• [ ] "-" Dash
• [ ] "_" Underscore
• [ ] "+" Plus
• [ ] "~" Tilde
• [ ] " " Space
• [ ] "Nothing" i.e., just remove those characters

[GitMurf]

@phhq made a good point:

The "/" is an important character -- to preserve a user's ability to use the namespacing features in Roam, eg. I used the excel functions like TEXTJOIN to format the first column to import names with namespaces. To accomplish a mass search and replace after being converted to markdown, the character(s) replacing the "/" should be very likley NOT to occur in natural filenames (and also not be on the "exclude" list.

Leaning towards:

• / --> ~
  • Due to important need for namespaces
• : --> +
• " --> '
• ? --> ^
• | --> ^^
• The following I don't see a big use case for really needing for any particular reason and would just be inclined to replace them all with _ underscore.
  • \ --> _
  • * --> _
  • < --> _
  • > --> _

[GitMurf]

Also I just thought of an idea... it is a little more "hacky/manual" but if you are going to do manual find/replace within Roam for some of this stuff, especially namespaces, then this is probably actually better:

• For any pages that have an illegal character... or maybe just the important/common ones like / and : and |
• Instead of creating pages for those ones, you instead just nest underneath their [[]] page name in the Summary csv-import sheet I create. Then Roam will create the page as you want it with / or : etc. based off the link when you import.
• Then you just have to look at the Summary page and see if there are any that have children bullets under it because those were the ones we had to skip .MD file creation and their attributes are just in child bullets underneath it on Summary page
• So then you simply copy/cut the attributes from under it, click the page link that was created during import, and paste in the page!!!

Thoughts?? I think I LOVE this idea and much better workaround. Thoughts? Am I not thinking about something?