Invalid UTF-8 JSON error while restoring large file about couchdb-dump HOT 5 CLOSED

I'd recommend loading the JSON in a JSON-friendly IDE, or sticking it through an online JSON validator to find out where the issue is.
My initial thinking is that you seem to have strange HTML character database names... What other strange stuff might be in that dump that needs escaping 😅

from couchdb-dump.

Well, the dump seems to be correct, I can open and validate it in VSCode. If I increase lines to 40000 it is imported correctly, so the problem is definitely a problem with the split.
DB name is also correct (it has some escaped chars).
Do you know how the split files are created normally? Should they be valid JSON files or not?

from couchdb-dump.

Yes, split files should be valid JSON, as they're loaded back into the DB during the restore process file-by-file (rather than being stitched back and inserted as a huge amount for example).

from couchdb-dump.

I found the reason for the issue. Basically, I was post-processing dump with jq and it was pretty-printing the file, causing couchdb-dump to get lost when splitting it in chunks. Thank you for your help in any case 😉

from couchdb-dump.

All good then 😁

from couchdb-dump.