0

I want nearly the same as answered here for pandas - but want to run it in vaex.
As vaex does lazy copy, for me it would be okay, to save (my two) columns of str.split into the vaex-df. But there is nothing like expand=True.

Bastian Ebeling
  • 1,138
  • 11
  • 38

1 Answers1

1

To do this with vaex 4.0.0 you can't retrieve a token by using the slice directly, you have to use the slice inside the apply method.

Here is the example adapted for vaex:

import pandas as pd
import vaex

df = pd.DataFrame({'ticker' : ['spx 5/25/2001 p500', 'spx 5/25/2001 p600', 'spx 5/25/2001 p700']})

df_vaex = vaex.from_pandas(df)

df_vaex["ticker"].str.split(" ").apply(lambda x: x[-1])

And it gives the expected output:

Expression = lambda_function(str_split(ticker, ' '))
Length: 3 dtype: string (expression)
------------------------------------
0  p500
1  p600
2  p700