Hey @pixie79, thank you for taking the time to look into the code and make this PR! We are really happy to see the community growing! I took a look at the code and I have a couple of questions regarding the changes.

I see that you added support for parsing field_name: TYPE and field_name TYPE (without the colon). In the current version of the SDK we can only parse the one with the colon.

The fetching of the table structure is done in the method fetch_table_info where a request is sent to a uc endpoint and we get back a response which is parsed into TableInfo instance. TableInfo looks like this:

{
  "name": "table_name",
  "catalog_name": "...",
  "schema_name": "...",
  "columns": [
    {
      "name": "...",
      "type_text": "...",
      ...
    }
  ],
  ...
}

From what I’ve seen, columns and type_text always uses colons. So I’m not sure we need to handle the space format? Of course, if you ran into some examples that didn't work, please share them with us.
I tried to make a table with the SQL script you provided in the nested_example and then generate files with the current version of the tool and everything worked as expected.

Other than that, the example looks great and we can definitely keep it. Let me know what you think.

Hi

not at my computer at the moment. But yes only made the PR as when I fetched the proto schema using the tool from my table via UC it failed to work and errored. This was the fix that let it work correctly

Can you perhaps share the error that you got? And the steps on how to get a repro. I can't seem to reproduce the issue on my side currently.

Hey @pixie79, just following up on this PR, if you happen to have the repro steps, that would help us evaluate the fix.