Stata 11 help for mf_runningsum

help mata runningsum() -------------------------------------------------------------------------------

Title

[M-5] runningsum() -- Running sum of vector

Syntax

numeric vector runningsum(numeric vector x [, missing])

numeric vector quadrunningsum(numeric vector x [, missing])

void _runningsum(y, numeric vector x [, missing])

void _quadrunningsum(y, numeric vector x [, missing])

where optional argument missing is a real scalar that determines how missing values in x are treated:

1. Specifying missing as 0 is equivalent to not specifying the argument; missing values in x are treated as contributing 0 to the sum.

2. Specifying missing as 1 specifies that missing values in x are to be treated as missing values and turn the sum to missing.

Description

runningsum(x) returns a vector of the same dimension as x containing the running sum of x. Missing values are treated as contributing zero to the sum.

runningsum(x, missing) does the same but lets you specify how missing values are treated. runningsum(x, 0) is the same as runningsum(x). runningsum(x, 1) specifies that missing values are to turn the sum to missing where they occur.

quadrunningsum(x) and quadrunningsum(x, missing) do the same but perform the accumulation in quad precision.

_runningsum(y, x [, missing]) and _quadrunningsum(y, x [, missing]) work the same way, except that rather than returning the running-sum vector, they store the result in y. This method is slightly more efficient when y is a view.

Remarks

The running sum of (1, 2, 3) is (1, 3, 6).

All functions return the same type as the argument, real if argument is real, complex if complex.

Conformability

runningsum(x, missing), quadrunningsum(x, missing): x: r x 1 or 1 x c missing: 1 x 1 (optional) result: r x 1 or 1 x c

_runningsum(y, x, missing), _quadrunningsum(y, x, missing): input: x: r x 1 or 1 x c y r x 1 or 1 x c (contents irrelevant) missing: 1 x 1 (optional) output: y: r x 1 or 1 x c

Diagnostics

If missing = 0, missing values are treated as contributing zero to the sum; they do not turn the sum to missing. Otherwise, missing values turn the sum to missing.

_runningsum(y, x, missing) and _quadrunningsum(y, x, missing) abort with error if y is not p-conformable with x and of the same eltype. The contents of y are irrelevant.

Source code

Functions are built in.

Also see

Manual: [M-5] runningsum()

Help: [M-5] sum(); [M-4] mathematical, [M-4] utility


© Copyright 1996–2009 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index