ValueError when using .rows on sheet with hyperlinks

I have a workbook containing a worksheet with hyperlinks. When loading the workbook (use_iterators=True and data_only=True) and reading rows of sheet called "Modification Summary", I get a ValueError.

I can confirm this with the following code:

from openpyxl import load_workbook
wb = load_workbook("Issues/bug328.xlsx", read_only=True, data_only=True)
ws = wb['Modification Summary']
for row in ws.rows:
    for cell in row:
        print(cell.value)

The traceback is

Traceback (most recent call last):
  File "/Applications/WingIDE.app/Contents/MacOS/src/debug/tserver/_sandbox.py", line 1, in <module>
    # Used internally for debug sandbox under external interpreter
  File "/Users/charlieclark/Projects/openpyxl/openpyxl/worksheet/iter_worksheet.py", line 186, in get_squared_range
    retrieved_columns = dict([(c.column, c) for c in cells])
  File "/Users/charlieclark/Projects/openpyxl/openpyxl/worksheet/iter_worksheet.py", line 186, in <listcomp>
    retrieved_columns = dict([(c.column, c) for c in cells])
  File "/Users/charlieclark/Projects/openpyxl/openpyxl/worksheet/iter_worksheet.py", line 222, in get_cells
    value, data_type, style_id)
  File "/Users/charlieclark/Projects/openpyxl/openpyxl/cell/read_only.py", line 20, in __init__
    self._set_value(value)
  File "/Users/charlieclark/Projects/openpyxl/openpyxl/cell/read_only.py", line 92, in _set_value
    value = float(value)
builtins.ValueError: could not convert string to float:

Doesn't happen with when not read_only.

looked at the issue, it's because we assume that a formula always gives back a numerical value, while it could (in this case) be a link with no "cache" to cast to int/float. I'm working on a fix, just have to figure out how to correct the assumption without breaking something else in the process ;-)

@Eric Gazoni in such cases it should probably return None.

It's worth noting that this file cannot be validated.

Standard cells use a regex to check for numerical values in such cases. This can probably be added to the the ReadOnlyCell as el.FindText('v') will always return an empty string for empty tags and only None if there is no such tag.

Merged in ericgazoni/openpyxl-328/2.1 (pull request #28)

when reading a sheet with data_only, containing an hyperlink but no cache value, do not try to force it as a numerical value, just return None (fixes #328)

→ <<cset 8dbd1b3dea3c>>

when reading a sheet with data_only, containing an hyperlink but no cache value, do not try to force it as a numerical value, just return None (fixes #328)

→ <<cset d1e1ea5bcb2f>>

in data only mode, when reading an empty cell with no type, force value to be None (fixes #328)

→ <<cset 6512e1702db1>>

Merged in ericgazoni/openpyxl-328/2.1 (pull request #28)

when reading a sheet with data_only, containing an hyperlink but no cache value, do not try to force it as a numerical value, just return None (fixes #328)

→ <<cset 8dbd1b3dea3c>>

Removing version: 2.1.x (automated comment)