Raise exception on non-unique column index in to_hdf for fixed format. #7788
Conversation
|
hmm, maybe make this a |
|
Well, with the current fixed format it is in fact invalid. There is no way to get this right without changing |
|
it could be implemented (it just isn't). You are preventing the writing (which invalidates the reading) either way. The exception will be more informative ( Actually you might want to have a nice exception message saying that you CAN store this in 'table' format. |
| @@ -2680,6 +2680,9 @@ def write(self, obj, **kwargs): | |||
|
|
|||
| self.attrs.ndim = data.ndim | |||
| for i, ax in enumerate(data.axes): | |||
| if i == 0: | |||
| if not ax.is_unique: | |||
| raise ValueError("Columns index has to be unique for fixed format") | |||
There was a problem hiding this comment.
say, you can store this in table format, however.
|
It could not be implemented because If this is the only thing keeping you from merging I'll change it to |
Raise exception on non-unique column index in to_hdf for fixed format.
|
@filmor thanks for this! |
|
this is technically for a non-unique info_axis (e.g. columns for df, but 0 axis for Series/Panel). Can you add some tests for this? And update the note in v0.15.0 (I moved it a bit) |
Fixes #7761.