GROUP_PREVIOUS

GROUP_PREVIOUS(<any>;[<number>])

Value: series value

K: lag

Returns the value of the set k'th previous record in series. This can be used for introducing a lag for time series analysis and/or computing autocorrelation.

This is a group series function.

Given the following data:

First create a group using GROUPBY(#RawData!Groups).

Then use the GROUP_PREVIOUS(#RawData!Number). The result shows the previous record's value in relation to the GROUPBY() column.

The following example illustrates how to compute the time a user has spent on a web site.

Given the following data:

First create a group using GROUPBY(#RawData!Session).

Next sort your timestamp in ascending order using GROUP_SORT_ASC(#RawData!TimeStamp)

Next use the GROUP_PATH(#RawData!url). The result shows paths for all values of the column in ascending order in relation to the GROUPBY() column.

Next use GROUP_PREVIOUS(#RawData!TimeStamp)

Finally use DIFF(#Sheet1!Ascending_TimeStamp;#Sheet1!Previous_TimeStamp) to subtract the ascending TimeStamp from the Previous_TimeStamp to find how long the user spend on the URL.