Command Reference : Operator and Function Reference : Cumulative Statistic Functions
  
Cumulative Statistic Functions
These functions perform basic running or cumulative statistics for a series and may be used as part of a series expression. The functions are split into two types, those that cumulate forwards and those that cumulate backwards. The forwards cumulating functions return the running values of a statistic from the start of the workfile (or optionally a sample) to the current observation. The backwards cumulating functions return the running values from the end of the workfile (or sample) to the current observation.
By default, EViews will use the entire workfile range when computing the statistics. You may provide the optional sample s as a literal (quoted) sample expression or a named sample.
Missing values, NAs, do not propagate through these functions. Thus the cumulative sums of the numbers 1, 3, 4, NA, 5 are 1, 4, 8, 8, 13.
 
Name
Function
Description
@cumsum(x[,s])
cumulative sum
cumulative sum of the values in X from the start of the workfile/sample.
@cumprod(x[,s])
cumulative product
cumulative product of the values in X from the start of the workfile/sample (note this function could be subject to numerical overflows).
@cummean(x[,s])
cumulative mean
mean of the values in X from the start of the workfile/sample to the current observation.
@cumstdev(x[,s])
cumulative standard deviation
sample standard deviation of the values in X from the start of the workfile/sample to the current observation. Note this calculation involves division by .
@cumstdevp(x[,s])
cumulative population standard deviation
population standard deviation of the values in X from the start of the workfile/sample to the current observation. Note this calculation involves division by .
@cumstdevs(x[,s])
cumulative sample standard deviation
sample standard deviation of the values in X from the start of the workfile/sample. Note this performs the same calculation as @cumstdev.
@cumvar(x[,s])
cumulative variance
population variance of the values in X from the start of the workfile/sample to the current observation. Note this calculation involves division by .
@cumvarp(x[,s])
cumulative population variance
population variance of the values in X from the start of the workfile/sample to the current observation. Note this performs the same calculation as @cumvar.
@cumvars(x[,s])
cumulative sample variance
sample variance of the values in X from the start of the workfile/sample to the current observation. Note this calculation involves division by .
@cummax(x[,s])
cumulative maximum
maximum of the values in X from the start of the workfile/sample to the current observation.
@cummin(x[,s])
cumulative minimum
minimum of the values in X from the start of the workfile/sample to the current observation.
@cumsumsq(x[,s])
cumulative sum-of-squares
sum of squares of the values in X from the start of the workfile/sample to the current observation.
@cumobs(x[,s])
cumulative nmber of non-NA observations
the number of non-missing observations in X from the start of the workfile/sample to the current observation.
@cumnas(x[,s])
cumulative number of NA observations
the number of missing observations in X from the start of the workfile/sample to the current observation.
@cumquantile(x,q[,s])
cumulative quantile
the qth quantile in X computed from the start of the workfile/sample to the current observation. Quantiles are computed using the Cleveland definition.
@cummedian(x[,s])
cumulative median
the median of X computed from the end of the workfile/sample to the current observation.
@cumbsum(x[,s])
backwards cumulative sum
cumulative sum of the values in X from the end of the workfile/sample.
@cumbprod(x[,s])
backwards cumulative product
cumulative product of the values in X from the end of the workfile/sample (note this function could be subject to numerical overflows).
@cumbmean(x[,s])
backwards cumulative mean
mean of the values in X from the end of the workfile/sample to the current observation.
@cumbstdev(x[,s])
backwards cumulative standard deviation
sample standard deviation of the values in X from the end of the workfile/sample to the current observation. Note this calculation involves division by .
@cumbstdevp(x[,s])
backwards cumulative population standard deviation
population standard deviation of the values in X from the end of the workfile/sample to the current observation. Note this calculation involves division by .
@cumbstdevs(x[,s])
backwards cumulative sample standard deviation
sample standard deviation of the values in X from the end of the workfile/sample. Note this performs the same calculation as @cumstdev.
@cumbvar(x[,s])
backwards cumulative variance
population variance of the values in X from the end of the workfile/sample to the current observation. Note this calculation involves division by
@cumbvarp(x[,s])
backwards cumulative population variance
population variance of the values in X from the end of the workfile/sample to the current observation. Note this performs the same calculation as @cumvar.
@cumbvars(x[,s])
backwards cumulative sample variance
sample variance of the values in X from the end of the workfile/sample to the current observation. Note this calculation involves division by .
@cumbmax(x[,s])
backwards cumulative maximum
maximum of the values in X from the end of the workfile/sample to the current observation.
@cumbmin(x[,s])
backwards cumulative minimum
minimum of the values in X from the end of the workfile/sample to the current observation.
@cumbsumsq(x[,s])
backwards cumulative sum-of-squares
sum of squares of the values in X from the start of the workfile/sample to the current observation.
@cumbobs(x[,s])
backwards cumulative nmber of non-NA observations
the number of non-missing observations in X from the end of the workfile/sample to the current observation.
@cumbnas(x[,s])
backwards cumulative nmber of NA observations
the number of missing observations in X from the end of the workfile/sample to the current observation.
@cumbquantile(x,q[,s])
backwards cumulative quantile
the qth quantile in X computed from the end of the workfile/sample to the current observation. Quantiles are computed using the Cleveland definition.
@cumbmedian(x[,s])
backwards cumulative median
the median of X computed from the end of the workfile/sample to the current observation.