Unusual IDs Preceding Graph Names


I have been a user of Logseq on my previous Manjaro installation, successfully managing several synchronized graphs across devices. The graph names followed a neat pattern such as “Logseq/Name1,” “Logseq/Name2,” and so forth. I assumed the “Logseq/” prefix was derived from the container folder name.

Recently, I switched to a new Ubuntu machine and opted for the Flatpack version of Logseq due to some challenges with the snapd version, but this is another issue. I’ve noticed a change in the naming convention for the graphs; they now appear as something like “3dda7ac654/Name1,” “de6a9ea4df/Name2,” and so on.

While the functionality remains intact, the new naming convention appears ugly. I recall on my Manjaro machine, initially creating a single graph under “Logseq” and later, when necessary, creating new graphs under it for specific content. I believe I might have removed the directories generated by Logseq for each graph directly under “Logseq.”

I would appreciate any insights or suggestions on how to revert to the previous naming convention or tidy up the current one. Thank you in advance for your assistance.



  • Could try to:
    • take a backup
    • restore the exact folder structure
    • re-index
    • hope for the best
  • Otherwise:
    • Don’t use Logseq as the name of a folder.
    • Treat Logseq as part of the system (i.e. protected/hidden), not as a container (and thus folder).
      • Therefore, move each graph outside folder Logseq and where it has more affinity with. That also:
        • prevents some single-point-of-failure scenarios
        • makes for better comparisons and overall organization
    • Name your graphs as domains and your pages as concepts.
    • Keep in mind than in future versions most data will be kept in a database, simplifying the file system.
1 Like

Awesome advice. Thank you!

And Happy New Year!

Thank you, mentaloid, for your valuable advice!

I’d like to provide further clarity on some aspects from my previous messages.

In my initial Logseq installation on my previous computer, I began with a single graph. During the setup, the application prompted me to choose a local directory for storage.

As my Logseq usage evolved, I found it advantageous to create distinct graphs for various activities. For organizational purposes, I allocated each graph its directory. All these directories were nested under the original “Logseq” directory. But I’m almost sure that after organizing the new graphs under the original Logseq directory there was no residue of the original graph version at that level.

I didn’t intentionally implement any namespacing strategy for the graph names. It’s importan to note that the names chosen for the graphs didn’t include an attempt at namespacing. For instance, a graph meant for personal content was simply named “Personal” and not “Logseq/Personal.”

However, despite this, the graph selector in the top left corner displayed them in a “Logseq/Name1” format, a choice I didn’t make at all.

In my new installation on the same machine but with a different operating system, I accessed each of the previous graphs stored on another disk within the same computer.

Interestingly, the graph selector now exhibits the graphs with an alphanumeric identifier instead of the original “Logseq/NameN” format.

Furthermore, I want to emphasize that I don’t require any namespace in the selector. In fact, I would prefer to use plain names without the alphanumeric identifier, and also without the “Logseq/” prefix.

I welcome any insights or suggestions on how to revert to the previous namespacing or address this alteration in the graph identifiers.

Thank you for your valuable assistance.