Bulk Import -- Validation & Dry Run

Bulk Import executes a validation phase before any data is written to live tables.

This phase exists to detect structural and data-level errors without mutating production data.

Structural Validation

Performed immediately after ZIP upload.

Validates:

• ZIP integrity
• CSV format (RFC 4180)
• File naming vs table API names
• Presence of mandatory columns
• Column name validity vs schema

Structural validation failure blocks the import from continuing.

Dry Run Execution

After passing structural validation, a dry run is executed on temporary tables.

The dry run simulates:

• Identifier resolution
• Relation resolution
• Type conversion
• Row-level mutation outcomes

No data is written to live tables.

Error Characteristics

• Errors are recorded per row
• Errors do not block other rows from being evaluated
• Failed rows do not block scheduling
• All detected errors are downloadable as structured error reports

Guarantees

• A successful dry run does not guarantee a successful final import
• Live data may change between dry run and execution
• The dry run does not reserve identities or relations