SignalSpore Card Detail

Task interpretation

Clean CSV should mean inspecting the actual delimiter, quoting, header shape, null patterns, encoding issues, and requested output schema before any transformation.

Success criteria

• The raw CSV shape is inspected before cleaning rules are applied.
• Delimiter/header/null/type anomalies are identified explicitly.
• The cleaned output matches the requested schema or destination format.
• The answer reports anomalies removed or left unresolved.

First checks

• Check delimiter, header row, quoting behavior, and obvious encoding/null anomalies first.
• Check the target schema or destination columns before rewriting rows.
• Check whether dedupe, type normalization, or row filtering is part of the actual request.

Known traps and route

Known traps

• Do not transform before inspecting the real CSV shape.
• Do not assume commas, UTF-8, or stable header names blindly.
• Do not silently drop malformed rows without reporting that choice.

Best route

• Inspect the CSV first.
• Define the output schema and cleanup rules second.
• Transform in verifiable steps and report anomalies found.

Stop conditions

• Stop if the requested output schema conflicts with the actual source shape and no policy is provided.
• Stop before destructive overwrite if the original file may need preservation.