maxframe.liteframe.LiteFrame.expanding#

LiteFrame.expanding(min_periods=1, columns=None, order_by=None, order_by_ascending=True, window_shift=0, reverse=False)#

Provide expanding window calculations.

Returns a LiteFrameExpanding object whose aggregation methods (.sum(), .mean(), …) compute a growing cumulative window from the start to the current row. Unlike cum*, expanding fills NaN positions with the accumulated value.

Parameters:

min_periods (int, default 1) – Minimum number of observations in the window required to produce a value (otherwise null).
columns (str, list of str, or None, default None) – Column(s) to aggregate. If None, all columns are aggregated. Non-selected columns are passed through unchanged.
order_by (str, list of str, or None, default None) – Column(s) that define the order in which the window expands (equivalent to ORDER BY inside a MaxCompute OVER() clause). The output row order is unchanged; each row receives the aggregation computed over the rows that precede it in order_by order.
order_by_ascending (bool or list of bool, default True) – Sort direction for each order_by column. A single bool is broadcast to all columns.
window_shift (int, default 0) – Offset the leading edge of the growing window by N rows. A positive value extends the window forward (the window becomes [0, i+shift]), a negative value shrinks it. Matches the MaxCompute shift MF-specific window parameter.
reverse (bool, default False) – Flip the expanding window so it grows from the last row backward (a “narrowing” window covering [i, n-1]). Matches the MaxCompute reverse_range MF-specific window parameter. May be combined with window_shift: the window then covers [i+window_shift, n-1].

Returns:

Intermediate object; call an aggregation method to produce a LiteFrame.

Return type:

LiteFrameExpanding