Concatenation

Coral.Frame.concat stacks frames vertically. Pass frames with the same column names and types and a shared static row count. The result has the rows of each input in list order.

concat takes its output columns and their order from the first frame, then looks up each of those columns by name in every later frame:

Later frame	Result
Same columns in a different order	Accepted; values are matched by name, and the first frame's order is kept
An extra column	The column is omitted, without an error
Missing a first-frame column	Fails with missing column: NAME
A first-frame column with a different type	Fails with the accessor error for the first frame's type, such as column is not int: NAME

An empty list returns a frame with no columns and zero rows.

module Coral.BookConcat
import Coral.Frame (BoolCol, from_pairs, concat, nrows, int_col_of_list)
export (main)
def main() -> i64 = {
  lhs = from_pairs([("id", int_col_of_list([1i64, 2i64])), ("flag", BoolCol(to_tensor([true, false])))])
  rhs = from_pairs([("id", int_col_of_list([3i64, 4i64])), ("flag", BoolCol(to_tensor([true, true])))])
  nrows(concat([lhs, rhs]))
}

main returns 4. concat accepts IntCol, FloatCol, StringCol, and BoolCol; it carries integer missing-value masks through to the result. A different row count can be rejected because all inputs share the same Frame[n] type in this function's signature.