rowlens

rowlens is a streaming CSV search CLI for very large files. It scans row by row, keeps memory usage flat, and prints matches in a readable terminal report with column names beside each value.

Why it exists

When a CSV is 17 GB, the usual "open it in a spreadsheet" workflow is already dead. rowlens is built for the needle-in-a-haystack case:

• Match one or more exact cell values with repeated --keyword
• Require one or more substring filters with repeated --filter
• Stream the file without loading it into memory
• Print each matching row in a bordered, columnar CLI layout
• Optionally save the same report to a text file with --output

Install

poetry install

Usage

poetry run rowlens --file "data.csv" --keyword "1213131" --filter "AAA"

poetry run rowlens \
  --file "huge.csv" \
  --keyword "1213131" \
  --keyword "YYYY" \
  --filter "AAA" \
  --filter "needle" \
  --ignore-case \
  --max-results 5 \
  --output "output/results.txt"

Matching rules

• Every --keyword must match a full cell value in the row.
• Every --filter must appear as a substring in at least one cell in the row.
• A row is returned only if it satisfies all supplied criteria.
• Reported row numbers are CSV record numbers, so the header is row 1 and the first data row is row 2.

Options

--file         CSV file to search
--keyword      Exact cell value to match; repeatable
--filter       Substring to require; repeatable
--delimiter    CSV delimiter, default ","
--encoding     File encoding, default "utf-8-sig"
--ignore-case  Case-insensitive matching
--max-results  Stop after N matches
--output       Write the rendered report to a text file

Packaging

The project uses Poetry and exposes a console script named rowlens, so it is ready to publish to PyPI after you set the final author metadata and repository URLs in pyproject.toml.

Tests

poetry run pytest