[polars] Fetch the FRED CSV with requests before parsing it

[mmcky]

Problem

The weekly cache build has failed since polars 2.0.0 was released on 2026-10-06. It passed on 2026-10-05 and failed on 2026-10-12. polars.md hits the 600 s cell timeout on this line:

data = pl.read_csv(fred_url, try_parse_dates=True)

The lecture runs !pip install --upgrade polars yfinance, so CI gets the newest polars. The zh-cn edition fails in the same place, and its execution report names this cell.

Cause

polars 2.0 reads a URL through its object store. It sends a HEAD request for the file size, then ranged GETs. FRED's fredgraph.csv answers HEAD with content-length: 0, although the body is 13,809 bytes. Ranged GETs work (206, content-range: bytes 0-3/13809). polars 2.0 then stalls instead of raising.

Run against the real URL:

polars	pl.read_csv(fred_url, try_parse_dates=True)
1.44.2	ok, 918 rows, 0.23 s
2.0.0	still running when killed after 90 s

Fix

Download the CSV with requests and pass the bytes to pl.read_csv. This gives 918 rows with the observation_date column parsed as Date under both 1.44.2 and 2.0.0. The sentence above the cell said read_csv fetches the URL itself, so it now describes the download.

This also matches the approach the pandas lecture already takes in its FRED section.

Testing

I ran all 52 code cells of the patched lecture under polars 2.0.0 and yfinance 1.7.0, with live FRED and Yahoo Finance data. All passed. The GitHub-hosted test_pwt.csv cells, which are untouched, also read fine under 2.0.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Wyb6EzbNzY5a8uuJPnDtUb

---

Generated by Claude Code

[Copilot]

🟡 Changes recommended

The fix is correct, but the newly added import requests is placed mid-cell, deviating from the lecture's convention of grouping imports at the top.

1 open finding

• Move requests import to top-level imports cell · New

What changed in this PR

This PR fixes the weekly cache build, which has been failing since polars 2.0.0 (released 2026-10-06). In the polars lecture, pl.read_csv(fred_url, ...) relies on polars reading the URL directly through its object store; under polars 2.0, FRED's fredgraph.csv returns content-length: 0 on HEAD, causing the read to stall until the 600 s cell timeout. The fix downloads the CSV with requests first and passes the bytes to pl.read_csv, which aligns with how the pandas lecture already fetches FRED data.

Changes:

• Download the FRED CSV via requests.get(...) and pass r.content to pl.read_csv instead of letting polars fetch the URL.
• Update the surrounding prose so it describes the requests download rather than read_csv fetching the URL.

File	Description
lectures/​polars.md	Fetch the FRED CSV with requests and pass the bytes to pl.read_csv; update the explanatory sentence accordingly.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[github-actions]

🚀 Deployed on https://6acdb0645ec78f853d2cfaf4--epic-agnesi-957267.netlify.app

[mmcky]

✅ Translation sync completed (zh-cn)

Target repo: QuantEcon/lecture-python-programming.zh-cn
Translation PR: QuantEcon/lecture-python-programming.zh-cn#108
Files synced (1):

• lectures/polars.md

[mmcky]

✅ Translation sync completed (fr)

Target repo: QuantEcon/lecture-python-programming.fr
Translation PR: QuantEcon/lecture-python-programming.fr#97
Files synced (1):

• lectures/polars.md

[mmcky]

✅ Translation sync completed (fa)

Target repo: QuantEcon/lecture-python-programming.fa
Translation PR: QuantEcon/lecture-python-programming.fa#171
Files synced (1):

• lectures/polars.md