Computes a non-overlapping block maximum of a series.
BLOCKMAX(series, N, "naflag")
series |
- |
A series or table. |
||||||
N |
- |
An integer, the points per block to calculate the maximum as the series is processed. |
||||||
"naflag" |
- |
Optional. A string, the NA handling method.
|
A series or table, the N-point block maximum.
W1: 1..12
W2: blockmax(W1, 2)
W2 == {2, 4, 6, 8, 10, 12}
W1: 1..12
W2: blockmax(W1, 3)
W2 == {3, 6, 9, 12}
W1: {1, 2, 3, 4, 5, nan, 7, 8, 9, nan, 11, 12}
W2: blockmax(W1, 3)
W3: blockmax(W1, 3, "includenan")
W2 == {3, 6, 9, 12}
W3 == {3, nan, 9, nan}
W2 ignores the NaN values.
W3 returns NaN for any block that includes one or more NaN values.
BLOCKMAX computes the N-point non-overlapping block maximum of a series by decimating the standard N-point moving maximum.
If the block size is not an integral fraction of the series length, the end points are not processed.
NaN values are ignored by default. Set naflag to "includenan" to process NaN values.
See XBLOCKMAX to compute the moving block maximum based on a X duration.