Fix for reading tdms_index files with extra padding after metadata (#197

)
adamreeve · May 5, 2020 · 3d4ba3e · 3d4ba3e
1 parent 560f7b7
commit 3d4ba3e
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 2 deletions.
diff --git a/nptdms/reader.py b/nptdms/reader.py
@@ -75,6 +75,7 @@ def read_metadata(self):
                 # Read metadata first to work out how much space we need
                 previous_segment = None
                 while True:
+                    start_position = file.tell()
                     try:
                         segment = self._read_segment_metadata(
                             file, segment_position, previous_segment, reading_index_file)
@@ -88,8 +89,11 @@ def read_metadata(self):
                     previous_segment = segment
 
                     segment_position = segment.next_segment_pos
-                    if not reading_index_file:
-                        file.seek(segment.next_segment_pos)
+                    if reading_index_file:
+                        lead_size = 7 * 4
+                        file.seek(start_position + lead_size + segment.raw_data_offset, os.SEEK_SET)
+                    else:
+                        file.seek(segment.next_segment_pos, os.SEEK_SET)
         finally:
             if reading_index_file:
                 file.close()

diff --git a/nptdms/test/scenarios.py b/nptdms/test/scenarios.py
@@ -974,6 +974,29 @@ def channel_without_data_or_data_type():
     return test_file, expected_data
 
 
+@scenario
+def extra_padding_after_metadata():
+    test_file = GeneratedFile()
+    test_file.add_segment(
+        ("kTocMetaData", "kTocRawData", "kTocNewObjList"),
+        segment_objects_metadata(
+            channel_metadata("/'group'/'channel1'", TDS_TYPE_INT32, 2),
+        ) + "00 00 00 00 00 00 00 00",
+        "01 00 00 00" "02 00 00 00"
+    )
+    test_file.add_segment(
+        ("kTocMetaData", "kTocRawData", "kTocNewObjList"),
+        segment_objects_metadata(
+            channel_metadata("/'group'/'channel1'", TDS_TYPE_INT32, 2),
+        ) + "00 00 00 00 00 00 00 00",
+        "03 00 00 00" "04 00 00 00"
+    )
+    expected_data = {
+        ('group', 'channel1'): np.array([1, 2, 3, 4], dtype=np.int32),
+    }
+    return test_file, expected_data
+
+
 def timestamp_data_chunk(times):
     epoch = np.datetime64('1904-01-01T00:00:00')