From abb61a0c3684aebc6f675e6be522048e5881f226 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ond=C5=99ej=20C=C3=ADfka?= <cifkao@users.noreply.github.com>
Date: Thu, 22 Jan 2026 18:25:33 +0100
Subject: [PATCH 1/3] Load RTTM with `keep_default_na=False`

---
 src/pyannote/database/util.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/pyannote/database/util.py b/src/pyannote/database/util.py
index 07d4bcf..e1e670d 100644
--- a/src/pyannote/database/util.py
+++ b/src/pyannote/database/util.py
@@ -180,7 +180,7 @@ def load_rttm(file_rttm, keep_type="SPEAKER"):
         names=names,
         dtype=dtype,
         sep=r"\s+",
-        keep_default_na=True,
+        keep_default_na=False,
     )
 
     annotations = dict()

From a83c483d98092294d09aef5b337162de6786ed48 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ond=C5=99ej=20C=C3=ADfka?= <cifkao@users.noreply.github.com>
Date: Thu, 22 Jan 2026 19:02:24 +0100
Subject: [PATCH 2/3] Add handling for NA values in non-float columns

---
 src/pyannote/database/util.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/pyannote/database/util.py b/src/pyannote/database/util.py
index e1e670d..7297168 100644
--- a/src/pyannote/database/util.py
+++ b/src/pyannote/database/util.py
@@ -175,11 +175,13 @@ def load_rttm(file_rttm, keep_type="SPEAKER"):
         "NA6",
     ]
     dtype = {"uri": str, "start": float, "duration": float, "speaker": str}
+    na_values = {n: ["<NA>"] for n in names if dtype.get(n) is not str}
     data = pd.read_csv(
         file_rttm,
         names=names,
         dtype=dtype,
         sep=r"\s+",
+        na_values=na_values,
         keep_default_na=False,
     )
 

From 1d5c784119c313e21216c7456e0578bfd27dac3b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ond=C5=99ej=20C=C3=ADfka?= <cifkao@users.noreply.github.com>
Date: Thu, 22 Jan 2026 19:20:48 +0100
Subject: [PATCH 3/3] Allow <NA> for all but replace with None for string
 columns

---
 src/pyannote/database/util.py | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/src/pyannote/database/util.py b/src/pyannote/database/util.py
index 7297168..95bb026 100644
--- a/src/pyannote/database/util.py
+++ b/src/pyannote/database/util.py
@@ -175,24 +175,25 @@ def load_rttm(file_rttm, keep_type="SPEAKER"):
         "NA6",
     ]
     dtype = {"uri": str, "start": float, "duration": float, "speaker": str}
-    na_values = {n: ["<NA>"] for n in names if dtype.get(n) is not str}
     data = pd.read_csv(
         file_rttm,
         names=names,
         dtype=dtype,
         sep=r"\s+",
-        na_values=na_values,
+        na_values=["<NA>"],
         keep_default_na=False,
     )
 
     annotations = dict()
     for uri, turns in data.groupby("uri"):
+        if pd.isna(uri):
+            uri = None
         annotation = Annotation(uri=uri)
         for i, turn in turns.iterrows():
             if turn.type != keep_type:
                 continue
             segment = Segment(turn.start, turn.start + turn.duration)
-            annotation[segment, i] = turn.speaker
+            annotation[segment, i] = turn.speaker if not pd.isna(turn.speaker) else None
         annotations[uri] = annotation
 
     return annotations