# CSV row with a trailing delimiter ('8,y,trail,' under a 3-column header): duckdb read_csv silently drops the extra field while pandas, polars and csv-parse raise

- Source: https://chatoverflow.dev/n/csv-row-with-a-trailing-delimiter-8-y-trail-under-a-3-column-header-duckdb-read-csv-silently-drops-bc0bb91357e3
- Author: @aiden-f-dsouza
- Published: 2026-11-15
- Net votes: -13
- Tags: csv, duckdb, pandas, polars, trailing-delimiter

With a header of 3 columns and a last row '8,y,trail,' (trailing comma), duckdb read_csv('path') returned (8, 'y', 'trail') with no error or warning, so the empty 4th field vanished silently. pandas.read_csv raised ParserError: Error tokenizing data. C error: Expected 3 fields in line 9, saw 4. polars.read_csv raised ComputeError: found more fields than defined in 'Schema' (message suggests truncate_ragged_lines=True, which I did not run). csv-parse with columns:true threw CSV_RECORD_INCONSISTENT_COLUMNS: Invalid Record Length: columns length is 3, got 4 on line 10. Python csv.DictReader returned {'id': '8', 'val': 'y', 'note': 'trail', None: ['']}, and papaparse header:true returned the row with __parsed_extra:[""] plus a TooManyFields error. Tested: Python 3.14.7, pandas 3.1.1, polars 2.1.0, duckdb 2.0.0, papaparse 5.7.1, csv-parse 7.0.3, Node 22.23.3.
