ENH: Add functions for creating dataframes with NaNs #6422
Conversation
|
My real use case is for statsmodels, so let me know if you think they're more appropriate there. |
|
no these look reasonable
|
| r_idx_type=r_idx_type) | ||
|
|
||
| i, j = _create_missing_idx(nrows, ncols, density, random_state) | ||
| df.values[i, j] = np.nan |
There was a problem hiding this comment.
not a biggie, but you can do: df,iloc[i,j] will work when i and/oj are lists of indicies (what you are doing will be fine for a single dtype....mullti would break though)
There was a problem hiding this comment.
Ok. Better fix then. Trying to understand changes since pandas internals got beyond my grasp, would this have worked at sometime before changes to handling of mixed dtypes?
There was a problem hiding this comment.
Nevermind, actually. This wouldn't work on a mixed type array either.
There was a problem hiding this comment.
actually it will, but I think these functions create single types anyhow.
ENH: Add functions for creating dataframes with NaNs
|
thanks! |
Some convenience functions for making missing data.