Skip to content

Conversation

@brycekbargar
Copy link
Collaborator

Contains fixes for bugs reporting by users switching to the 3.1 release.

Keys in JSON aren't ordered, so far LDLite has been implicitly relying on the first key in the response to be the one that has the items. This has worked so far but starting in Sunflower keys seem a little weirder. Now we're iterating through all the known "not a list of records" keys when trying to locate the actual key.
During various refactors the example script stopped running and it isn't something that is regularly checked. This makes sure it continues running after catching errors and surfaces those errors more clearly. It also fixes valid syntax that doesn't actually work at runtime.

After having troubles with duckdb, the example is being switched to sqlite.

Some people have also had trouble after installing ldlite because of the postgres dependency. This adds additional instructions to the readme for it.
When scanning the data for column types, currently the "last" record wins and overwrites any other datatype that has been found. I'm not quite sure why it iterates over every record this way but I'm trying not to touch the existing transform code unless absolutely necessary.

This change checks the datatype this column was previously parsed as and then doesn't allow a type to be "upgraded" from a varchar to something more specific because that will cause issues inserting the actual data later.
@brycekbargar brycekbargar merged commit 20cb96b into main Dec 8, 2025
7 of 9 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

2 participants