Replacing values in a data frame

Specific Domains Data
1

Don’t know if this is the right category for this…

I have a dataframe with a column containing averages calculated with the mean function. These values go out several decimal places, but I want to use round() to set it to two or three. Aside from just having populated the dataframe with rounded values to begin with, how can I iterate through the column replacing each value with it’s rounded version? I’ve tried several different ways, .=, .==, replace(), but nothing is replacing the values. Any suggestions?
Thank you

2

Eg some variation on

@. df.somecol = round(df.somecol; digits = 2)
2 Likes
3

Thank you very much. I’m pretty new to programming and Julia. I wonder what does the @ do compared to what I had done?
Thank you

4

Without using @., the following should also do it:
df.somecol = round.(df.somecol; digits = 2)
df.somecol is a vector, and the . broadcasts the round function to apply to each element of the vector.

5

Note that

@. df.somecol = round(df.somecol; digits = 2)

is not equivalent to

df.somecol = round.(df.somecol; digits = 2)

but to

df.somecol .= round.(df.somecol; digits = 2)
1 Like
6

Correct, but in this context (as far as I can see), .= does the same as = because both sides already have the same length due to broadcasting of round.

7

If he is using DataFrames.jl I think that = will replace the old vector by a new one, while .= will alter the old vector object inplace (and because broadcast fusing this will probably not allocate memory).

1 Like
8

Thanks, I did not know this!
In my toy example, replacing = by .= gave a speed-up of nearly a factor 2.

1 Like