NAG CL Interface
g01apc (quantiles_​stream_​arbitrary)

1 Purpose

g01apc finds approximate quantiles from a large arbitrary-sized data stream using an out-of-core algorithm.

2 Specification

#include <nag.h>
void  g01apc (Integer *ind, const double rv[], Integer nb, double eps, Integer *np, const double q[], double qv[], Integer nq, double rcomm[], Integer lrcomm, Integer icomm[], Integer licomm, NagError *fail)
The function may be called by the names: g01apc, nag_stat_quantiles_stream_arbitrary or nag_approx_quantiles_arbitrary.

3 Description

A quantile is a value which divides a frequency distribution such that there is a given proportion of data values below the quantile. For example, the median of a dataset is the 0.5 quantile because half the values are less than or equal to it.
g01apc uses a slightly modified version of an algorithm described in a paper by Zhang and Wang (2007) to determine ε-approximate quantiles of a large arbitrary-sized data stream of real values, where ε is a user-defined approximation factor. Let m denote the number of data elements processed so far then, given any quantile q0.0,1.0, an ε-approximate quantile is defined as an element in the data stream whose rank falls within q-εm,q+εm . In case of more than one ε-approximate quantile being available, the one closest to qm is used.

4 References

Zhang Q and Wang W (2007) A fast algorithm for approximate quantiles in high speed data streams Proceedings of the 19th International Conference on Scientific and Statistical Database Management IEEE Computer Society 29

5 Arguments

1: ind Integer * Input/Output
On initial entry: must be set to 0.
On entry: indicates the action required in the current call to g01apc.
ind=0
Initialize the communication arrays and attempt to process the first nb values from the data stream. eps, rv and nb must be set and licomm must be at least 10.
ind=1
Attempt to process the next block of nb values from the data stream. The calling program must update rv and (if required) nb, and re-enter g01apc with all other parameters unchanged.
ind=2
Continue calculation following the reallocation of either or both of the communication arrays rcomm and icomm.
ind=3
Calculate the nq ε-approximate quantiles specified in q. The calling program must set q and nq and re-enter g01apc with all other parameters unchanged. This option can be chosen only when npexp1.0/eps.
On exit: indicates output from the call.
ind=1
g01apc has processed np data points and expects to be called again with additional data.
ind=2
Either one or more of the communication arrays rcomm and icomm is too small. The new minimum lengths of rcomm and icomm have been returned in icomm[0] and icomm[1] respectively. If the new minimum length is greater than the current length then the corresponding communication array needs to be reallocated, its contents preserved and g01apc called again with all other parameters unchanged.
If there is more data to be processed, it is recommended that lrcomm and licomm are made significantly bigger than the minimum to limit the number of reallocations.
ind=3
g01apc has returned the requested ε-approximate quantiles in qv. These quantiles are based on np data points.
Constraint: ind=0, 1, 2 or 3.
2: rv[dim] const double Input
Note: the dimension, dim, of the array rv must be at least
  • nb, when ind=0,1 or 2.
On entry: if ind=0, 1 or 2, the vector containing the current block of data, otherwise rv is not referenced.
3: nb Integer Input
On entry: if ind=0, 1 or 2, the size of the current block of data. The size of blocks of data in array rv can vary; therefore nb can change between calls to g01apc.
Constraint: if ind=0, 1 or 2, nb>0.
4: eps double Input
On entry: approximation factor ε.
Constraint: eps>0.0 ​ and ​eps1.0.
5: np Integer * Output
On exit: m, the number of elements processed so far.
6: q[dim] const double Input
Note: the dimension, dim, of the array q must be at least
  • nq, when ind=3.
On entry: if ind=3, the quantiles to be calculated, otherwise q is not referenced. Note that q[i]=0.0, corresponds to the minimum value and q[i]=1.0 to the maximum value.
Constraint: if ind=3, 0.0q[i-1]1.0, for i=1,2,,nq.
7: qv[dim] double Output
Note: the dimension, dim, of the array qv must be at least
  • nq, when ind=3.
On exit: if ind=3, qv[i] contains the ε-approximate quantiles specified by the value provided in q[i].
8: nq Integer Input
On entry: if ind=3, the number of quantiles requested, otherwise nq is not referenced.
Constraint: if ind=3, nq>0.
9: rcomm[lrcomm] double Communication Array
On entry: if ind=1 or 2 then the first l elements of rcomm as supplied to g01apc must be identical to the first l elements of rcomm returned from the last call to g01apc, where l is the value of lrcomm used in the last call. In other words, the contents of rcomm must not be altered between calls to this function. If rcomm needs to be reallocated then its contents must be preserved. If ind=0 then rcomm need not be set.
On exit: rcomm holds information required by subsequent calls to g01apc
10: lrcomm Integer Input
On entry: the dimension of the array rcomm.
Constraints:
  • if ind=0, lrcomm1;
  • otherwise lrcommicomm[0].
11: icomm[licomm] Integer Communication Array
On entry: if ind=1 or 2 then the first l elements of icomm as supplied to g01apc must be identical to the first l elements of icomm returned from the last call to g01apc, where l is the value of licomm used in the last call. In other words, the contents of icomm must not be altered between calls to this function. If icomm needs to be reallocated then its contents must be preserved. If ind=0 then icomm need not be set.
On exit: icomm[0] holds the minimum required length for rcomm and icomm[1] holds the minimum required length for icomm. The remaining elements of icomm are used for communication between subsequent calls to g01apc.
12: licomm Integer Input
On entry: the dimension of the array icomm.
Constraints:
  • if ind=0, licomm10;
  • otherwise licommicomm[1].
13: fail NagError * Input/Output
The NAG error argument (see Section 7 in the Introduction to the NAG Library CL Interface).

6 Error Indicators and Warnings

NE_ALLOC_FAIL
Dynamic memory allocation failed.
See Section 3.1.2 in the Introduction to the NAG Library CL Interface for further information.
NE_ARRAY_SIZE
On entry, licomm=value.
Constraint: licomm10.
On entry, lrcomm=value.
Constraint: lrcomm1.
NE_BAD_PARAM
On entry, argument value had an illegal value.
NE_ILLEGAL_COMM
The contents of icomm have been altered between calls to this function.
The contents of rcomm have been altered between calls to this function.
NE_INT
On entry, ind=0, 1 or 2 and nb=value.
Constraint: if ind=0, 1 or 2 then nb>0.
On entry, ind=3 and nq=value.
Constraint: if ind=3 then nq>0.
On entry, ind=value.
Constraint: ind=0, 1, 2 or 3.
NE_INTERNAL_ERROR
An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
See Section 7.5 in the Introduction to the NAG Library CL Interface for further information.
NE_NO_LICENCE
Your licence key may have expired or may not have been installed correctly.
See Section 8 in the Introduction to the NAG Library CL Interface for further information.
NE_Q_OUT_OF_RANGE
On entry, ind=3 and q[value]=value.
Constraint: if ind=3 then 0.0q[i]1.0 for all i.
NE_REAL
On entry, eps=value.
Constraint: 0.0<eps1.0.
NE_TOO_SMALL
Number of data elements streamed, value is not sufficient for a quantile query when eps=value.
Supply more data or reprocess the data with a higher eps value.

7 Accuracy

Not applicable.

8 Parallelism and Performance

g01apc is threaded by NAG for parallel execution in multithreaded implementations of the NAG Library.
Please consult the X06 Chapter Introduction for information on how to control and interrogate the OpenMP environment used within this function. Please also consult the Users' Note for your implementation for any additional implementation-specific information.

9 Further Comments

The average time taken by g01apc scales as np log 1/ε logεnp .
It is not possible to determine in advance the final size of the communication arrays rcomm and icomm without knowing the size of the dataset. However, if a rough size (n) is known, the speed of the computation can be increased if the sizes of the communication arrays are not smaller than
lrcomm = log2 n×eps+1.0 - 2 × 1.0/eps +1+x+ 2× minx, x/2.0 +1 × y +1 licomm = log2 n×eps+1.0 - 2 × 2 × 1.0/eps +1 + 1 + 2 × x+2× minx, x/2.0 +1 × y + y + 11  
where
x= max1, log eps×n / eps y = log2n/x+1.0 +1 .  

10 Example

This example computes a list of ε-approximate quantiles. The data is processed in blocks of 20 observations at a time to simulate a situation in which the data is made available in a piecemeal fashion.

10.1 Program Text

Program Text (g01apce.c)

10.2 Program Data

Program Data (g01apce.d)

10.3 Program Results

Program Results (g01apce.r)