From 6100a867f67015d59ad80795bd481bf4820f78b4 Mon Sep 17 00:00:00 2001
From: Kevin Costner <120246174+kevincostner17@users.noreply.github.com>
Date: Tue, 15 Sep 2026 00:29:09 +0530
Subject: [PATCH] docs: changelog entry for the pandas < 2 nullable-int
 streaming fixes

#192 fixed a user-facing crash in fd.clean_timeseries / StreamingCleaner on
nullable integer columns with gaps under pandas < 2, and #195 applied the
same NA-safe conversion to MissForest, but neither added a CHANGELOG entry.
---
 CHANGELOG.md | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 67fb564d..3a2f4499 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -31,6 +31,11 @@ adheres to [Semantic Versioning](https://semver.org/).
 - The Polars engine's duplicate-detection and dedup row counts now run through
   the streaming collect path instead of re-evaluating the whole plan in memory
   (#53).
+- `fd.clean_timeseries` and `StreamingCleaner` no longer crash with
+  "cannot convert to 'float64'-dtype NumPy array with missing values" on
+  nullable integer columns (`Int*`/`UInt*`) with gaps under pandas < 2
+  (running statistics and short-gap interpolation). MissForest's convergence
+  check uses the same NA-safe conversion.
 
 ## [2.0.0] - 2026-07-20