AttributeError: 'float' object has no attribute 'strip'

[ruedi.silvestri@…]

I got successfully TicketImportPlugin 0.8 and xlrd 0.7.1 install. Though trying to upload a Excel 2003 format xls results to an "AttributeError: 'float' object has no attribute 'strip'" exception.

Datei "build/bdist.win32/egg/trac/web/main.py", Zeile 517, in _dispatch_request
Datei "build/bdist.win32/egg/trac/web/main.py", Zeile 238, in dispatch
Datei "build/bdist.win32/egg/talm_importer/importer.py", Zeile 77, in process_request
Datei "build/bdist.win32/egg/talm_importer/importer.py", Zeile 121, in _do_preview
Datei "build/bdist.win32/egg/talm_importer/importer.py", Zeile 348, in _process 

[ruedi.silvestri@…]

I just identified the issues being caused by having blanks in the file name I tried to upload. Maybe it would help to catch this exception with a respective error message as a workaround.

Thanks, Ruedi.

[ruedi.silvestri@…]

please ignore the above comment about the filename. I managed to identify the cause which was origined by a text value in a cell which is supposed to be a number. Actually it would be good to make below extensions:

1) validate format of source value and raise exceptions if issues found. 2) pure number values seem automatically to be considered as floating numbers. Though on some reference numbers one would not want to have the number suffixed by .0 by the import.

Regards, Ruedi.

[François Granade]

Ruedi,

Could you attach a .xls file that shows the problem to this ticket ? I don't have Excel 2003 and it will save me some time....

Thanks for reporting and for investigating !

[Jun Omae]

I attach Excel 2003 file which contains text, number, date, boolean and error cells, celltypes.xls​. When excel file has boolean and error cells other than number cells, the same issue occurs.

The following patch; convert the value to string for each cell type.

talm_importer/readers.py

@@ -102 +102 @@
             row = {}
             i = 0
             for cx in xrange(self.sh.ncols):
-                row[header[i]] = self.sh.cell_value(rx, cx)
-                if self.sh.cell_type(rx, cx) == xlrd.XL_CELL_DATE:
-                    row[header[i]] = datetime.datetime(*xlrd.xldate_as_tuple(row[header[i]], self.book.datemode)).strftime(self._datetime_format)
+                val = self.sh.cell_value(rx, cx)
+                cell_type = self.sh.cell_type(rx, cx)
+                if cell_type == xlrd.XL_CELL_NUMBER:
+                    val = '%g' % val
+                elif cell_type == xlrd.XL_CELL_DATE:
+                    val = datetime.datetime(*xlrd.xldate_as_tuple(val, self.book.datemode))
+                    val = val.strftime(self._datetime_format)
+                elif cell_type == xlrd.XL_CELL_BOOLEAN:
+                    val = ('FALSE', 'TRUE')[val]
+                elif cell_type == xlrd.XL_CELL_ERROR:
+                    val = xlrd.error_text_from_code.get(val) or '#ERR%d' % val
+                row[header[i]] = val
 
                 i += 1
             data.append(row)

[Jun Omae]

(In [10230]) fixed #8804, convert value of cell which is number, boolean or error cell to string.