setColumnarTuple!

Syntax

setColumnarTuple!(X, [on=true])

Details

Marks a tuple as a columnar tuple in place, or clears the columnar tuple flag.

A columnar tuple is still an ANY tuple composed of scalars or vectors, but DolphinDB interprets it using columnar rules similar to those of array vectors: each element in the tuple corresponds to one row value of a table column; if an element is a vector, that row stores an array. A regular tuple is interpreted by tuple rules, where each row is a complete ANY element when the tuple is used as a table column.

For example, when ungroup is applied to a table column containing vector elements, a regular tuple column is not flattened. After it is converted to a columnar tuple, ungroup flattens the vector elements in each row.

Parameters

X is a tuple composed of vectors or scalars of the same data type.

on (optional) is a Boolean value indicating whether to mark X as a columnar tuple. The default value is true.

If on = true, converts a regular tuple to a columnar tuple.
If on = false, clears the columnar tuple flag and converts it to a regular tuple.

Returns

Returns the modified X.

If on = true, returns a columnar tuple.
If on = false, returns a regular tuple.

Examples

Convert a regular tuple to a columnar tuple.

tp = [[1,2,3], [4,5,6], [7,8]]
isColumnarTuple(tp)
// output: false

tp.setColumnarTuple!()
isColumnarTuple(tp)
// output: true

When a regular tuple is used as a table column, each row is an ANY element. ungroup does not flatten the vector elements in it.

t = table([1,2] as id, ([10,11], 12) as v)
ungroup(t)


id	v
1	[10,11]
2	12

After the regular tuple is converted to a columnar tuple, ungroup flattens column v by array-vector rules.

t = table([1,2] as id, ([10,11], 12).setColumnarTuple!() as v)
ungroup(t)


id	v
1	10
1	11
2	12

Clear the columnar tuple flag.

v = ([10,11], 12).setColumnarTuple!()
isColumnarTuple(v)
// output: true

v.setColumnarTuple!(false)
isColumnarTuple(v)
// output: false