Can I effectively import block references from Roam Research so they link across pages?

I have been using Roam Research for more than a year, and have about 6,000 pages in RR. I have experimented with moving using MD and JSON exports from RR. Importing the JSON files using the import function, and saving the MD files into the Pages directory and re-indexing.

Two issues that I have not figured out how to handle and would like some help with.

  1. Importing block references - is there a way of getting logseq to recognise block references between pages? I used the Roam42 DB Stats that tells me I have over 69,000 references within my RR graph. I used it a lot to use the same version of information in multuple places. Is anyone able to help?

  2. And if that one proves possible, I have over 1,700 Firebase links and while these are imported into logseq successfully, I don’t want to leave those images in Firebase and the RR environment. I can’t find any advice on how to do this, but did find this link that is targeted at Obsidian users, but should work on the exported MD files from RR. Does anyone have a good approach to doing this?

1 Like

If you import the Roam JSON using the “Import Roam JSON” function in Logseq, it should recognize and reformat all of the block references/embeds, as well as date pages to work properly in Logseq (just did my own Roam db, 2 years, 9000 pages). Only queries need a little bit of polish to work.

For Firebase images, any script that can move through all the Markdown files, download the files, and then update the link, should work.


Thank you for the guidance. I had been importing parts of the graph separately so logseq wasn’t linking the blocks correctly. I tried a full export/import and, as you advised, all the links were created correctly.

I haven’t used queries much as I never really found a practical use case for them that I needed. My main issue is that I have used a lot of RR attributes, although I will just convert those to pages.

How long did this take? I have also 2yr, ~2k pages and it’s been all morning. Not sure if it’s hung up somewhere.

1 Like