Extract a Parsed Table Into Structured Records
Beginner
The problem, from first principles
A table extracted from a PDF or an image (by an OCR pass or a table-detection model) usually comes out as a header row plus a list of data rows -- both just lists of strings. Reasoning about that table by column name ("what's the price column") is much easier once each row is a dict keyed by header name, rather than a positional list the caller has to remember the column order of.
The task
Write rows_to_records(header, rows) that pairs each row with the header by position and returns one dict per row. A row shorter than the header simply omits its missing trailing columns; a row longer than the header drops its extra trailing cells.
Keywords
production-systemsmultimodaldata-processing
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.