Skapa nya kolumner

Datamanipulering i Julia

Katerina Zahradova

Instructor

Kolumner vs. rader

Kolumn

Ett enda värde för alla rader
T.ex. medelvärde, median, summa, ...

Rader

Värden beror på data i varje enskild rad

=> använd ByRow()

Fenans längd i tum

# Convert mm to inches
transform(penguins, :flipper_length_mm => ByRow(x -> x/25.4) => :flipper_length_inch)

333x8 DataFrame
Row species   island    ...  body_mass_g  sex      flipper_lenght_inch
    String15  String15  ...  Int64        String7  Float64 
___________________________________________________________________
1   Adelie    Torgersen ...  3750         MALE     7.12598
2   Adelie    Torgersen ...  3800         FEMALE   7.32283
...

Kvot mellan näbbdjup och näbblängd

# Select columns and calculate their ratio
select(penguins,:culmen_depth_mm, :culmen_length_mm, 
   [:culmen_depth_mm, :culmen_length_mm] => ByRow((x, y) -> x/y) => :culmen_ratio)

333x3 DataFrame
Row  culmen_depth_mm  culmen_length_mm  culmen_ratio
     Float64          Float64           Float64
_____________________________________________________
1    18.7             39.1              0.478261
2    17.4             39.5              0.440506
...

Ny kolumn från en vektor

# Vector id_vec that we want to add


# Using [] and :
penguins[:, :id_colon] = id_vec


# Using [] and !
penguins[!, :id_exclamation] = id_vec


# Using .
penguins.id_dot = id_vec

Kopia eller referens

penguins[:, :id_colon] = id_vec
penguins[!, :id_exclamation] = id_vec
penguins.id_dot = id_vec

# Change first element
id_vec[1] = 27
select(penguins, :species, r"id")

333x4 DataFrame
Row  species   id_colon  id_exclamation  id_dot
     String15  Int64     Int64           Int64
_____________________________________________________
1    Adelie    25        27              27
...

Kopia eller referens

penguins[:, :id_colon] = id_vec        # copies values to the DataFrame
penguins[!, :id_exclamation] = id_vec  # references id_vec
penguins.id_dot = id_vec               # references id_vec

# Change first element
id_vec[1] = 27
select(penguins, :species, r"id")

333x4 DataFrame
Row  species   id_colon  id_exclamation  id_dot
     String15  Int64     Int64           Int64
_____________________________________________________
1    Adelie    25        27              27
...

Nu kör vi en övning!

Datamanipulering i Julia