bug in DataFrame/DataMatrix toCSV method when index=False

When index=False an extra column is added because the leading comma is never removed.

```
def toCSV(self, path, nanRep='', cols=None, header=True, index=True, mode='wb'):

    *<snip>*

    for idx in self.index:

        if index:
            f.write(str(idx))
        for col in cols:
            *<snip>*
            f.write(',%s' % val)
```
