what is the normalization for json sources (like mongodb) to a relational database

[axellpadilla]

Hi and thanks ahead, searched for it, but couldn't find an answer to what sling does for column names and arrays on json objects, will it normalize and create extra tables if needed? will it clean names of columns?

[flarco]

Hi, by default it keeps the JSON as source, and does not flatten. If you set the source_options.flatten = true, then object keys will become their own columns (type inferred), and arrays will stay as it (as a JSON type column). Sling will not unnest.

[axellpadilla]

thank you, so also column names remain the same or is there an option to normalize?

[flarco]

Not sure what you mean by normalize, but there is the column_casing target option.

[axellpadilla]

that's close, thanks, normalize is a step more than casing (remove spaces, special symbols, etc)

[flarco]

ah yes, it does: https://github.com/slingdata-io/sling-cli/blob/main/core/sling/task.go#L540