Page History

...

We would like to propose 4 kinds of window TVF: tumbling, hopping, cumulative, session. The return value of window TVF is a relation that includes all columns of data as well as additional one column named window to indicate the assigned window. The window column is a composite ROW type consists of 3 columns named window_start, window_end and , window_time columns. The to indicate the assigned window. The window_time field is a time attribute of the record after window TVF, it alwsays equal to "window_end - 1". 

...

The table-valued function TUMBLE  assigns a window for each row of a relation based on a time attribute column. The return value of TUMBLE is a relation that includes all columns of data as well as additional a ROW type column named window to 3 columns named window_start, window_end, window_time to indicate the assigned window. The original row time attribute "timecol" will be a regular timestamp column after window TVF. All assigned windows have the same length, and that’s why tumbling sometimes is named as “fixed windowing”.

...

The table-valued function HOP assigns windows that cover rows within the interval of size and shifting every slide based on a timestamp column. The return value of HOP is a relation that includes all columns of data as well as additional a ROW type column named window additional 3 columns named window_start, window_end, window_time to indicate the assigned window. The original row time attribute "timecol" will be a regular timestamp column after window TVF. Windows assigned could have overlapping so hopping sometime is named as “sliding windowing”.

...

The table-valued function CUMULATE assigns windows that cover rows within an initial interval of step size, and expanding to one more step size (keep window start fixed) every step until to the max window size. The return value of CUMULATE is a relation that includes all columns of data as well as additional a ROW type column named window 3 columns named window_start, window_end, window_time to indicate the assigned window. The original row time attribute "timecol" will be a regular timestamp column after window TVF. Windows assigned could have overlapping.

...

The table-valued function SESSION assigns windows that cover rows based on datetime. Within a session window, distances of rows are less than interval. Session window is applied per key. The return value of SE is a relation that includes all columns of data as well as additional a ROW type column named window additional 3 columns named window_start, window_end, window_time to indicate the assigned window. The original row time attribute "timecol" will be a regular timestamp column after window TVF.  

...