better conditional expressions

[universalmind303]

Is your feature request related to a problem? Please describe.

conditional logic in daft is hard to write and read

# %%
import daft
# %%
data = {
  "age": [12, 20, 19, 30, 62]
}

df = daft.from_pydict(data).sort('age')

is_senior = daft.col('age') >=55)
is_adult = daft.col('age') >= 18)

df = df.with_column(
  "age_bracket",
  is_senior.if_else(
    "senior",
    is_adult.if_else("adult", "child")
  )
)


print(df.collect())

It does not have a natural flow to it, and it makes it difficult to build complex conditional logic.

Describe the solution you'd like

I think even a purely functional syntax would be better than the chaining syntax here.

condition(    # if
  is_senior,  # is_senior: 
  'senior',   # 'senior
  condition(  # else if
    is_adult, # is_adult:
    'adult',  # 'adult'
    'child'   # else: 'child
  )
)

you could easily add kwargs to this for more readability

condition(
  if = is_senior,
  then = 'senior',
  else = condition(
    if = is_adult,
    then = 'adult',
    else = 'child' 
  )
)

Alternatively, some languages use a when/then/otherwise syntax

is_senior = daft.col('age') >=55)
is_adult = daft.col('age') >= 18)

condition = (daft
    .when(is_senior).then('senior')
    .when(is_adult).then('adult')
    .otherwise('child'))
    
df = df.with_column('age_bracket' condition)

Other dsl's use a switch/match syntax

age = daft.col('age')

(daft.match(age)
  .case(age >= 55, 'senior')
  .case(age >= 18, 'adult')
  .case(19, 'nineteen')
  .case(20, 'twenty')
  .default('child'))
           

---

IMO, any of these proposed alternatives are more intuitive than the current syntax.