ODSReader: Parse numbers correctly from .ods files

#343

Author: Bjwebb

CHANGELOG.md

@@ -8,7 +8,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ### Fixed
 
-- Fix parsing number formatting from .ods files https://github.com/OpenDataServices/flatten-tool/pull/373
+- Fix parsing date and number formatting from .ods files https://github.com/OpenDataServices/flatten-tool/pull/373
 
 ## [0.15.3] - 2020-02-23
 

flattentool/ODSReader.py

@@ -22,7 +22,6 @@
 import odf.opendocument
 from odf.table import Table, TableCell, TableRow
 
-
 # Backport for datetime.fromisoformat, which is new in Python 3.7
 backports.datetime_fromisoformat.MonkeyPatch.patch_fromisoformat()
 
@@ -79,10 +78,16 @@ def readSheet(self, sheet):
                             )
                         )
                         if value_type == "float":
+                            value = cell.attributes.get(
+                                (
+                                    "urn:oasis:names:tc:opendocument:xmlns:office:1.0",
+                                    "value",
+                                )
+                            )
                             if "." in str(cell):
-                                arrCells[count] = float(str(cell))
+                                arrCells[count] = float(value)
                             else:
-                                arrCells[count] = int(str(cell))
+                                arrCells[count] = int(value)
                         elif value_type == "date":
                             date_value = cell.attributes.get(
                                 (

flattentool/tests/fixtures/ods/types.ods

@@ -193,13 +193,18 @@ def test_xlsx_input_types(self):
         assert list(xlsxinput.get_sheet_lines("main")) == [
             {
                 "colInt": 1,
-                "colFloat": 1.2,
+                "colFloat": 1000.2,
+                "colFloatComma": 1000.2,
                 "colDate": datetime.datetime(2020, 3, 5),
                 "colDateTime": datetime.datetime(2020, 2, 7, 16, 41, 0, 1),
+                None: None,
             }
         ]
         assert type(list(xlsxinput.get_sheet_lines("main"))[0]["colInt"]) == int
         assert type(list(xlsxinput.get_sheet_lines("main"))[0]["colFloat"]) == float
+        assert (
+            type(list(xlsxinput.get_sheet_lines("main"))[0]["colFloatComma"]) == float
+        )
         assert xlsxinput.sub_sheet_names == ["main"]
 
     def test_ods_input_types(self):
@@ -210,13 +215,14 @@ def test_ods_input_types(self):
         assert list(odsinput.get_sheet_lines("main")) == [
             {
                 "colInt": 1,
-                "colFloat": 1.2,
+                "colFloat": 1000.2,
+                "colFloatComma": 1000.2,
                 "colDate": datetime.datetime(2020, 3, 5),
                 "colDateTime": datetime.datetime(2020, 2, 7, 16, 41),
             }
         ]
         assert type(list(odsinput.get_sheet_lines("main"))[0]["colInt"]) == int
-        assert type(list(odsinput.get_sheet_lines("main"))[0]["colFloat"]) == float
+        assert type(list(odsinput.get_sheet_lines("main"))[0]["colFloatComma"]) == float
         assert list(odsinput.sub_sheet_names) == ["main"]
 
     def test_xlsx_input_integer2(self):