Export issues from 28.06 on with linked columns (Rows Endpoint)


I encountered an issue with the data export via Rows Endpoint for our data in case of specific columns which are linking the working table to a subtable. From 28.06 on we did not receive the data for the respective columns in multiple cases for our existing tables.

To recreate the issue I created a new TestDB with a table, a linked subtable.

After accessing the data with identical settings via the Rows-endpoint ( https://cloud.seatable.io/dtable-db/api/v1/query/ )

“convert_keys”: true,
“sql”: “SELECT * from Table1 LIMIT 2000”

we received the desired response – e.g. for row 3:

“Name”: “Max”,
“Verknüpfung zu anderen Einträgen”: [
“row_id”: “fR7yyHN4T9iNp2HFSrUlGQ”,
“display_value”: “Apple2”

The legacy data had the same structure prior to yesterday.
Now, however, we are getting an empty array as the JSON output instead (in multiple cases) in the legacy data.

Dear mv_ale,

I apologize for this problem.

I can confirm that the result of the List Rows (with SQL) sometimes don’t deliver the links. This situation might occur with “older bases”, meaning bases that were created with an older version of SeaTable. The new bases are working as expected. The developers will take a look at it and see if they can find the reason.

Here are ways to resolve the problem without waiting for a patch or update:

  • copy the base
  • export/import the base
  • make a snapshot, delete the link columns, recreate the link column and reassign the values

I know these workarounds are not perfect, but the best what I can deliver after 16h. Probably a fix will be available soon.

Maybe you have special operations causing dtable-db can’t handle the base correctly. You can try to add a new column to the base and then remove the column to make dtable-db refresh its metadata for the base, and then try calling the API again.