Documentation

Wraps a value into an Any type. This helps up represent rows as heterogenous lists.

Unwraps a value from an Any type. A Null cell yields Nothing.

Wrap a column cell into an Any, honouring the column's null bitmap: a slot marked invalid becomes Null, any other slot becomes a Value. Only needs Columnable a (the stored element type), not Columnable (Maybe a).

Reconstruct column i from a list of rows. The element type is taken from the first non-Null cell; a differently-typed cell is skipped. If any cell is Null the result is a nullable column, so a round-trip preserves nulls.

Convert the whole dataframe to a list of rows, one per row index in natural order; each row lists all columns ordered by column index. Materializes every row, so prefer toRowVector for large frames.

>>> toRowList df
[[("name", "Alice"), ("age", 25), ...], [("name", "Bob"), ("age", 30), ...], ...]

Convert the dataframe to a vector of rows containing only the named columns, in the given order. An empty name list yields one empty row per dataframe row.

>>> toRowVector ["name", "age"] df
Vector of rows with only name and age fields

Given a row gets the value associated with a field.

Examples

>>> map (rowValue (F.col @Int "age")) (toRowList df)
[25,30, ...]