From 1cfb6dee41dbd5624146c93f4d66d05873abd058 Mon Sep 17 00:00:00 2001
From: dsmedia <63077097+dsmedia@users.noreply.github.com>
Date: Sun, 19 Apr 2026 00:15:02 -0400
Subject: [PATCH] fix: declare pandas as runtime dep of parquet extra
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The ParquetParser unconditionally calls pyarrow.Table.to_pandas() and
constructs a TableResource(data=df, format="pandas"), so pandas is
required at runtime for any .parquet read. The `parquet` extra only
declared pyarrow, so `pip install 'frictionless[parquet]'` in isolation
raised ModuleNotFoundError: pandas on first read.

Tests never caught this because the hatch default env installs
frictionless[...,pandas,parquet,...] together, masking the gap. The
proper regression guard is a dedicated CI job that installs only
.[parquet] and runs the parquet tests in isolation — out of scope here.

- pyproject.toml: add pandas>=1.0 to the parquet extra
- CHANGELOG.md: note under [Unreleased]
---
 CHANGELOG.md   | 6 ++++++
 pyproject.toml | 2 +-
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 5a1b2c2f34..5ab800f81d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,12 @@
 
 Here described only the breaking and most significant changes. The full changelog and documentation for all released versions could be found in nicely formatted [commit history](https://github.com/frictionlessdata/frictionless-py/commits/main).
 
+## [Unreleased]
+
+- fix: declare `pandas` as a runtime dependency of the `parquet` extra — fixes
+  `ModuleNotFoundError: pandas` when reading `.parquet` files with only
+  `frictionless[parquet]` installed.
+
 ## v5.19.0
 
 - Deprecation warning for Python 3.8 and 3.9 ahead of future end of support
diff --git a/pyproject.toml b/pyproject.toml
index 02bcad8eb3..e613aa9e7e 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -93,7 +93,7 @@ html = ["pyquery>=1.4"]
 mysql = ["sqlalchemy>=1.4", "pymysql>=1.0"]
 ods = ["ezodf>=0.3", "lxml>=4.0"]
 pandas = ["pyarrow>=14.0", "pandas>=1.0"]
-parquet = ["pyarrow>=14.0"]
+parquet = ["pyarrow>=14.0", "pandas>=1.0"]
 postgresql = ["sqlalchemy>=1.4", "psycopg>=3.0", "psycopg2>=2.9"]
 spss = ["savReaderWriter>=3.0"]
 sql = ["sqlalchemy>=1.4"]