Feature Request: Add skip_rows_before_header (or equivalent) to CsvDecoder in Connector Builder

[ced455]

Summary
The CsvDecoder component in the Airbyte Connector Builder does not support skipping rows before the header row, blocking users from correctly parsing CSV/TSV files that contain metadata lines at the top.

---

Problem Description

When building a source connector using the Connector Builder against an API (Apple API in my case) that returns a TSV file, the response includes metadata on the first 3 lines before the actual header row. The CsvDecoder reads the first line as the header, resulting in incorrect schema detection and failed parsing.

---

Root Cause

The official CsvDecoder component only exposes three properties [YAML Reference]:

encoding (string) — default: utf-8
delimiter (string) — default: ,
set_values_to_none (array)

There is no parameter to skip rows before the header. This is in contrast to file-based source connectors (S3, GCS, Azure Blob Storage, etc.), which do support a skip_rows_before_header option [CSV format settings].

---

Impact / Blocked

Users building API-based connectors via the Connector Builder that receive CSV/TSV responses with metadata headers are completely blocked from correctly parsing those files. There is no workaround available within the Connector Builder.

---

Request

Add a skip_rows_before_header and skip_rows_after_header parameter to the CsvDecoder component, consistent with the behavior already available in file-based connectors, where it is defined as:

"The number of rows to skip before the header row. For example, if the header row is on the 3rd row, enter 2 in this field."

---

Proposed YAML Example

decoder:
type: CsvDecoder
delimiter: "\t"
encoding: "utf-8"
skip_rows_before_header: 3
skip_rows_after_header : 0

---

Priority Justification

Without this feature, users must resort to building a full custom Python connector, which is significantly more complex and has deployment restrictions on Airbyte Cloud.