Merge pull request #20 from heikomuller/dev-0.2.1

Fix bug when adding snapshot from file without primary key (#19)

changelog.md

@@ -39,3 +39,8 @@
 
 * Include wrapper for CSV files
 * Commit CSV files directly to a HISTORE archive
+
+
+### 0.2.1 - 2020-11-11
+
+* Fix bug when adding snapshot from file without primary key (\#19)

histore/archive/merge.py

@@ -86,6 +86,6 @@ def merge_rows(
         arch_row = arch_reader.next()
     # Add remaining document rows to the new archive version.
     while doc_row is not None:
-        # Outout an archive row created from the document row.
+        # Output an archive row created from the document row.
         writer.write_document_row(row=doc_row, version=version)
         doc_row = doc_reader.next()

histore/document/csv/read.py

@@ -133,6 +133,10 @@ def next(self):
         -------
         histore.document.row.DocumentRow
         """
+        # Return None if the reader is not set (i.e., because the end of the
+        # file has been reached).
+        if self.reader is None:
+            return None
         result = self._next_row
         try:
             rowid, row = next(self.reader)
@@ -151,9 +155,11 @@ def next(self):
                 values=values
             )
         except StopIteration:
-            # Close the file if the file is reached.
+            # Close the file if the end is reached. Set the reader to None in
+            # order to return None when next is called again.
             self._next_row = None
             self.reader.close()
+            self.reader = None
         # Return the buffered result
         return result
 

histore/version.py

@@ -6,4 +6,4 @@
 # file LICENSE for full license details.
 
 """Code version information for histore."""
-__version__ = '0.2.0'
+__version__ = '0.2.1'

tests/archive/store/test_persistent_archive.py

@@ -69,6 +69,20 @@ def test_persistent_archive(tmpdir):
     assert sorted([v.value for v in rows[0].cells[1].values]) == [32, 33]
 
 
+def test_merge_file_without_pk(tmpdir):
+    """Test merging snapshots of the NYC Watershed data into an archive without
+    using a primary key (issue #19).
+    """
+    archive = PersistentArchive(
+        basedir=str(tmpdir),
+        replace=True
+    )
+    s = archive.commit(WATERSHED_1)
+    diff = archive.diff(s.version - 1, s.version)
+    assert len(diff.schema().insert()) == 10
+    assert len(diff.rows().insert()) == 1793
+
+
 @pytest.mark.parametrize(
     'doc',
     [
@@ -77,12 +91,13 @@ def test_persistent_archive(tmpdir):
         WATERSHED_1
     ]
 )
-def test_watershed_archive(doc, tmpdir):
+@pytest.mark.parametrize('replace', [True, False])
+def test_watershed_archive(doc, replace, tmpdir):
     """Test merging snapshots of the NYC Watershed data into an archive."""
     archive = PersistentArchive(
         basedir=str(tmpdir),
         primary_key=['Site', 'Date'],
-        replace=False
+        replace=replace
     )
     s = archive.commit(doc)
     diff = archive.diff(s.version - 1, s.version)