Add a method to add schema information to a BigQueryRelation #1232
Conversation
| * @param schema The schema. | ||
| * @return A new relation with the schema added. | ||
| */ | ||
| public Relation addSchema(@Nullable Schema schema) { |
There was a problem hiding this comment.
Please validate there is no existing schema unless you expect schema overrides
There was a problem hiding this comment.
What should we ideally be doing in case a schema already exists? Since we're returning a new Relation I think overriding the schema should be fine.
There was a problem hiding this comment.
I don't know how do you plan to use this method. Generally it's better to introduce methods when they are needed (even with mock implementation) than visa versa.
There was a problem hiding this comment.
I've mentioned that already in the PR description above. I plan to use it in Wrangler in transform() just like any other Relation operation. I don't see any other way to add a schema to the Relation.
There was a problem hiding this comment.
Whangler should have nothing to do with schema management. How would Wrangler know the schema of a SQL expression?
There was a problem hiding this comment.
Also it's unneded burden on the plugin. Given known schema of the original relation and known schema of all expressions, resulting schema should be automatically contructed by the platform
DJSagarAhire
changed the title
This PR adds a way to add schema information to a relation. This can be called by a plugin that wishes to request validation.
The current intention is that this will be called by Wrangler in the
transform()method to supply schema information (which Wrangler currently already has in theoSchemavariable).