Delimited text format schema definition doesn't work

[vody]

@vishalaj1 creates an issue for documentation: https://github.com/MicrosoftDocs/azure-docs/issues/34997. Duplicating it in this repository as this looks like a bug in ADF rather then in documentation.

As per the syntax we can define schema.

"schema": [ < physical schema, optional, retrievable during authoring > ],

However, when i define schema: "schema": [ { "name": "Tst1", "type": "String" },

This does not work. The column names are still read as Prop_0 etc.

An expectation are:

• Order of fields in schema is matching column order in delimited text formatted file.
• Data Preview \ WranglingDataFlow \ Import Schema Functionality will use schema definition if this is possible. Logic should look like:
  • If "First row as header" is true then extract column names from file and match them with schema to get types
  • If "First row as header" is false then extract column names and types from schema based on column order
  • If schema is not defined then identify column as "Prop_{column number}" with undefined type. (current logic)

[hbzhxying]

Hi, vody

We don't support custom schema for delimited text. In UX, you cannot edit the schema also.

Thanks

[vody]

@hbzhxying, is this can be considered as a future request?

[hbzhxying]

@vody We will note down it and internal discuss about the plan.

Thanks

[vody]

@hbzhxying Our use case is: We are streaming data to Azure Data Lake v2 into a delimited text using Azure Blob Block per row line to enable row level updates. This was the smalless format for us which has been supported by ADF as well as by Power BI.