22  Adding rows

A late order arrives in a table of its own, so where do its rows go? add_rows stacks another table underneath this one. Nothing is matched and no column is added; the answer holds every row of both tables.

The late order is more, one row with the same columns as sales:

date	region	product	quantity	revenue	cost
2026-10-30	North	Sprocket	20	300	240

run

run('
  sales
    then add_rows more
    then sort [date] descending
    then take 3
')

date	region	product	quantity	revenue	cost
2026-10-30	North	Sprocket	20	300	240
2026-10-02	West	Widget	5	125	50
2026-09-15	East	Gadget	1	60	40

R

sales |> add_rows(more) |> sort(descending(date)) |> take(3)

date	region	product	quantity	revenue	cost
2026-10-30	North	Sprocket	20	300	240
2026-10-02	West	Widget	5	125	50
2026-09-15	East	Gadget	1	60	40

Python

sales >> add_rows(more) >> sort(descending(col.date)) >> take(3)

date	region	product	quantity	revenue	cost
2026-10-30	North	Sprocket	20	300	240
2026-10-02	West	Widget	5	125	50
2026-09-15	East	Gadget	1	60	40

sales then add_rows more then sort [date] descending then take 3

The sort is there so the page can show the stacking worked: newest first, and the late order is the first row of the answer.

The other table can even be this one, which is how a table is doubled:

run

run('
  sales
    then add_rows sales
    then summarize [rows] as row_count()
')

rows
30

R

sales |> add_rows(sales) |> summarize(rows = row_count())

rows
30

Python

sales >> add_rows(sales) >> summarize(rows = row_count())

rows
30

Every row appears twice, and both copies stay: add_rows never removes anything. Dropping repeats has a verb of its own, drop_duplicates.

Both tables need the same columns, and each column has to hold the same kind of thing on both sides. A column on one side only is refused rather than filled in with missing values, because a half-empty column that says nothing is how a mistake survives to the end of a pipeline.

R

late <- data.frame(region = "North")

collect(sales |> add_rows(late))

Error:
! 
illegal: `add_rows` needs both tables to have the same columns: this table has date, product, quantity, revenue, cost and `late` does not. Add what is missing, or drop it with `pick`
  |
2 |   then add_rows late
  |        ^^^^^^^^^^^^^

Python

late = pd.DataFrame({"region": ["North"]})

try:
    collect(sales >> add_rows(late))
except GodError as refusal:
    print(refusal)

illegal: `add_rows` needs both tables to have the same columns: this table has date, product, quantity, revenue, cost and `late` does not. Add what is missing, or drop it with `pick`
  |
2 |   then add_rows late
  |        ^^^^^^^^^^^^^

The message names the columns one side lacks, so the repair starts with the tables rather than with the sentence: make them agree, and the verb has nothing left to refuse.