Automatic conversion of stats from numeric to integer is causing loss of data

[patrickchodowski]

Automatic conversion of stats from numeric to integer is causing loss of data

[PatrickChodowski]

Hi, I removed the conversion to int completely, hopefully this helps and doesnt break anything

[PatrickChodowski]

Hello,

Thank you for your feedback! I will have a look

[PatrickChodowski]

Version 2.1.2 should fix the problem

[peteowen1]

Version 2.1.2 should fix the problem

Version 2.1.2 hasn't fixed the problem.
The problem seems to be when you use:

mutate_at(vars(- matches('_pct|spd|dist|_frequency|pie|per|_freq|_rating')), c_to_int)

the c_to_int function converts all these to an integer so a frequency of 0.04 gets converted to 0.

I don't get why c_to_int is needed at all??
Why do you need to convert numerics to integers can't we just leave them as numeric?

p.s. just recently stumbled upon the package and love your work, couldn't find a good scraper elsewhere and this package does everything I was looking for. Hoping by commenting I can contribute to it in someway!

[PatrickChodowski]

Hi, thanks for the feedback

originally with c_to_int I wanted to convert ID-like columns to integers (player_id, game_id, etc.) from floats

let me have a look again at it

[peteowen1]

Someone just opened up an issue with the same problem hahah

#6

I think I would just specify the ID columns to be integers and let all the other columns be numeric unless there's a good reason not to I suppose.
Especially since the default is per-game numbers which in general are always going to be a decimal therefore converting to integer will cause a loss of information.