Getting rid of user warning

[derrickcchow]

Hi,

When, I rename a column name. I am getting a UserWarning: The state wants to rename Partial Period to __Partial Period, but __Partial Period was not found, ignoring the rename. Is there a way to get rid of this error or even silence it?

Thanks, Derrick

[JovanVeljanoski]

Can you please give us minimal example? Also which vaex version are you using?

[derrickcchow]

vaex version: vaex~=4.7.0
I am using vaex with plotly dash components.

The column data type starts as a string.
df_vaex['Fiscal Week of Event'] = df_vaex.func.where(df_vaex['Fiscal Week of Event'] == "", np.nan, df_vaex['Fiscal Week of Event'])
df_vaex['Fiscal Week of Event'] = df_vaex['Fiscal Week of Event'].astype('float64')

[2022-01-26 15:04:23,022] INFO in _internal.py (fn:_log ln:113): 127.0.0.1 - - [26/Jan/2022 15:04:23] "POST /python/dashboard/_dash-update-component HTTP/1.1" 200 -
[2022-01-26 15:04:26,565] ERROR in config.py (fn:write ln:74): C:\Users\derek\AppData\Local\Programs\Python\Python39\lib\site-packages\vaex\dataframe.py:2628: UserWarning: The state wants to rename Week of Event to __Week of Event, but __Week of Event was not found, ignoring the rename
[2022-01-26 15:04:26,565] ERROR in config.py (fn:write ln:74): warnings.warn(f'The state wants to rename {old} to {new}, but {new} was not found, ignoring the rename')

[JovanVeljanoski]

I thought of a minimal example that we can run to reproduce your issue.
For instance I tried this:

import vaex
import numpy as np

df = vaex.from_arrays(x=['', '', '1', '2', '3'])
df.variables['nan'] = np.nan
df['x'] = df.func.where(df.x=='', np.nan, df.x)
df['x'] = df['x'].astype('float64')

And I don't get that user warning. Btw, are you getting just a warning or an actual error?

Another note:
This like df_vaex.func.where(df_vaex['Fiscal Week of Event'] == "", np.nan, df_vaex['Fiscal Week of Event']) leads to a dtype object, and is probably a bad idea..

@maartenbreddels maybe we should have a way to force-cast an expression to a numberic one, that will make all non-numeric elements np.nan. Basically similar to pd.to_numeric. There was also a request in the issues somewhere at some point.
What do you think?

Oh I found the issue, also @derrickcchow maybe this will be helpful to you too:
#1331

[derrickcchow]

It's more of an error message in the terminal!