Closed efstathios-chatzikyriakidis closed 4 months ago
Hi @avsolatorio!
Unfortunately, I still have a problem with the fix. I think that we need to do .astype('int64') instead of .astype(int). It is safer to convert datetime64[ns] to int64 because in some systems bare int could translate to int32.
Line that needs to be changed, from:
series = (series.astype(int) / 1e9)
To:
series = (series.astype('int64') / 1e9)
The fix will allow to run it everywhere with latest pandas.
Hi @avsolatorio,
I hope you are well.
Unfortunately, I am blocked in a situation where I need to upgrade pandas to latest (2.2.1) but I can't because the library REalTabFormer can't work with it. It seems that the function
process_datetime_data()
fails. Pandas has deprecated theSeries.view()
function, and we get both a warning and an error from that line that uses it:https://github.com/worldbank/REaLTabFormer/blob/4d14472f181f26b68b528630c372c6d828d7aa1e/src/realtabformer/data_utils.py#L259
The good thing is that it is the only place in the code that we use Series.view() so it might be easy to fix.
Can you help me on this? I will need a new PyPI version also (1.0.6). Thanks anyway.
WARNING:
ERROR:
Here is the list of python packages I am using: