If-else applied to a DataFrame

New to Julia
1

How can the IF ELSE loop be applied to a data frame.
I want the loop to take the first 2 columns as conditions and give me an output in the final.

For eg: if “col A” == 4 && “col B” == 2
print(“Yes”)
else
print(“no”)

2

Please check this out.

1 Like
3

A column of a dataframe is just a vector, so

for (a, b) in zip(df.A, df.B)
    (a==4) & (b==2) ? print("Yes") : print("No")
end
3 Likes
4

Or using broadcasting:

f(a,b) = (a==4) & (b==2) ? "Yes" : "No"
res = f.(df.A, df.B)
println(res)
2 Likes
5

How do you interpret the function broadcast?

6

df.A and df.B are Arrays with the same length. The function f is applied to each “row” of the dataframe separately and its output is written into the array res.
Hope this answers your question!

7

Does (df.A, df.B) == zip(df.A, df.B)?

8

No.
zip(df.A, df.B) gives an array of (2-element) tuples, whereas (df.A, df.B) is a tuple of arrays.

9

So f.(df.A, df.B) applies f to every pair of arguments that it extracts from each row of columns df.A and df.B and returns a vector of “yes” or “no” strings?

10

Correct

Example:

df = DataFrame(a=(1:10).+2, b=1:10)
f(a,b) = (a==4) & (b==2) ? "Yes" : "No"
res = f.(df.a, df.b)
11

A few more details: the . after a function is syntactic sugar. All dotted functions on a line are fused into a single function and applied element-wise to their arguments using the higher-level function broadcast. So

sin.(cos.(xs) .+ 1) == broadcast(x -> sin(cos(x) + 1), xs)

The function broadcast is similar to map, but it broadcasts singleton array dimensions, e.g.

julia> ones(10,2) .+ (1:10)
10×2 Array{Float64,2}:
  2.0   2.0
  3.0   3.0
  4.0   4.0
  5.0   5.0
  6.0   6.0
  7.0   7.0
  8.0   8.0
  9.0   9.0
 10.0  10.0
 11.0  11.0
3 Likes
12

Strange nobody mentioned ifelse.

julia> new_vector = ifelse.((df["col_A"] .== 4) .& (df["col_B"] .== 2), "Yes", "No")
julia> println(new_vector)
3 Likes