Unable to add columns to Table resource

[mattfysh]

What happened?

Attempting to define a table with the nodejs SDK as:

new databricks.Table('name', {
  // ...
  columns: [{
    name: 'x',
    position: 0,
    typeName: 'string',
    typeText: 'varchar(64)',
  }],
})

This results in the following error:

cannot update table: UpdateTable Missing required field: UpdateTable Missing required field: column_0.type_name

Expected Behavior

The plugin should correctly map typeName to type_name before sending to the databricks API

Steps to reproduce

Create any databricks.Table resource with the nodejs Pulumi SDK

Output of pulumi about

CLI
Version 3.55.0
Go Version go1.19.5
Go Compiler gc

Plugins
NAME VERSION
nodejs unknown

Host
OS darwin
Version 13.2.1
Arch x86_64

This project is written in nodejs: executable='/usr/local/bin/node' version='v19.7.0'

Additional context

No response

Contributing

Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[mattfysh]

Perhaps this is not possible? I am migrating from AWS Glue Catalog, which allows you to create a Table without any compute resources (since a catalog should be just metadata anyway)

But after playing with the Databricks console a little more, it appears that you need to use Compute resources to create the table so a new delta log can be written to S3

It's a shame it works this way, Spark jobs don't require any existing files at the table location when they first run, and this seems like a step away from making catalogs "metadata only", which they should be. in theory, metadata catalogs should not require any compute resources to create new schemas, tables and views.

[iwahbe]

Hey @mattfysh, thanks for opening an issue! I'm not sure I understand how your comment relates to this issue. Can you clarify what might be impossible?

[mjeffryes]

Closing due to inactivity.