Feature Type
-
[X] Adding new functionality to pandas
-
[ ] Changing existing functionality in pandas
-
[ ] Removing existing functionality in pandas
Problem Description
I have a very long and wide string dataframe (e.g. 4 million rows * 1000 columns) that need to be to_numericed.
The df contains invalid values that should be coerced as NaN. Currently the only approach is to use apply
df.apply(pd.to_numeric, errors="coerce")
which takes a substantial amount time given the number of columns.
Feature Description
Allow pd.to_numeric to take dataframe directly.
Alternative Solutions
Alternatively, if astype can take coerce, if can also solve the issue.
Additional Context
No response
Comment From: joooeey
I've reported this in #48781 .
Comment From: phofl
There is no real advantage of having to_numeric taking a DataFrame, we would still have to iterate over all columns, e.g. same outcome. Closing as there is another issue about astype already