 PDL::Ufunc(3) primitive ufunc operations for pdl

## DESCRIPTION

This module provides some primitive and useful functions defined using PDL::PP based on functionality of what are sometimes called ufuncs (for example NumPY and Mathematica talk about these). It collects all the functions generally used to "reduce" or "accumulate" along a dimension. These all do their job across the first dimension but by using the slicing functions you can do it on any dimension.

The PDL::Reduce module provides an alternative interface to many of the functions in this module.

use PDL::Ufunc;

## prodover

```  Signature: (a(n); int+ [o]b())
```

Project via product to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the product along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = prodover(\$a);
```

``` \$spectrum = prodover \$image->xchg(0,1)
```

prodover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## dprodover

```  Signature: (a(n); double [o]b())
```

Project via product to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the product along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = dprodover(\$a);
```

``` \$spectrum = dprodover \$image->xchg(0,1)
```

Unlike prodover, the calculations are performed in double precision.

dprodover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## cumuprodover

```  Signature: (a(n); int+ [o]b(n))
```

Cumulative product

This function calculates the cumulative product along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

The sum is started so that the first element in the cumulative product is the first element of the parameter.

``` \$b = cumuprodover(\$a);
```

``` \$spectrum = cumuprodover \$image->xchg(0,1)
```

cumuprodover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## dcumuprodover

```  Signature: (a(n); double [o]b(n))
```

Cumulative product

This function calculates the cumulative product along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

The sum is started so that the first element in the cumulative product is the first element of the parameter.

``` \$b = cumuprodover(\$a);
```

``` \$spectrum = cumuprodover \$image->xchg(0,1)
```

Unlike cumuprodover, the calculations are performed in double precision.

dcumuprodover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## sumover

```  Signature: (a(n); int+ [o]b())
```

Project via sum to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the sum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = sumover(\$a);
```

``` \$spectrum = sumover \$image->xchg(0,1)
```

sumover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## dsumover

```  Signature: (a(n); double [o]b())
```

Project via sum to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the sum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = dsumover(\$a);
```

``` \$spectrum = dsumover \$image->xchg(0,1)
```

Unlike sumover, the calculations are performed in double precision.

dsumover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## cumusumover

```  Signature: (a(n); int+ [o]b(n))
```

Cumulative sum

This function calculates the cumulative sum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

The sum is started so that the first element in the cumulative sum is the first element of the parameter.

``` \$b = cumusumover(\$a);
```

``` \$spectrum = cumusumover \$image->xchg(0,1)
```

cumusumover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## dcumusumover

```  Signature: (a(n); double [o]b(n))
```

Cumulative sum

This function calculates the cumulative sum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

The sum is started so that the first element in the cumulative sum is the first element of the parameter.

``` \$b = cumusumover(\$a);
```

``` \$spectrum = cumusumover \$image->xchg(0,1)
```

Unlike cumusumover, the calculations are performed in double precision.

dcumusumover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## andover

```  Signature: (a(n); int+ [o]b())
```

Project via and to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the and along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = andover(\$a);
```

``` \$spectrum = andover \$image->xchg(0,1)
```

If "a()" contains only bad data (and its bad flag is set), "b()" is set bad. Otherwise "b()" will have its bad flag cleared, as it will not contain any bad values.

## bandover

```  Signature: (a(n);  [o]b())
```

Project via bitwise and to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the bitwise and along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = bandover(\$a);
```

``` \$spectrum = bandover \$image->xchg(0,1)
```

If "a()" contains only bad data (and its bad flag is set), "b()" is set bad. Otherwise "b()" will have its bad flag cleared, as it will not contain any bad values.

## borover

```  Signature: (a(n);  [o]b())
```

Project via bitwise or to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the bitwise or along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = borover(\$a);
```

``` \$spectrum = borover \$image->xchg(0,1)
```

If "a()" contains only bad data (and its bad flag is set), "b()" is set bad. Otherwise "b()" will have its bad flag cleared, as it will not contain any bad values.

## orover

```  Signature: (a(n); int+ [o]b())
```

Project via or to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the or along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = orover(\$a);
```

``` \$spectrum = orover \$image->xchg(0,1)
```

If "a()" contains only bad data (and its bad flag is set), "b()" is set bad. Otherwise "b()" will have its bad flag cleared, as it will not contain any bad values.

## zcover

```  Signature: (a(n); int+ [o]b())
```

Project via == 0 to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the == 0 along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = zcover(\$a);
```

``` \$spectrum = zcover \$image->xchg(0,1)
```

If "a()" contains only bad data (and its bad flag is set), "b()" is set bad. Otherwise "b()" will have its bad flag cleared, as it will not contain any bad values.

## intover

```  Signature: (a(n); float+ [o]b())
```

Project via integral to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the integral along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = intover(\$a);
```

``` \$spectrum = intover \$image->xchg(0,1)
```

Notes:

"intover" uses a point spacing of one (i.e., delta-h==1). You will need to scale the result to correct for the true point delta).

For "n > 3", these are all "O(h^4)" (like Simpson's rule), but are integrals between the end points assuming the pdl gives values just at these centres: for such `functions', sumover is correct to O(h), but is the natural (and correct) choice for binned data, of course.

intover ignores the bad-value flag of the input piddles. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## average

```  Signature: (a(n); int+ [o]b())
```

Project via average to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the average along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = average(\$a);
```

``` \$spectrum = average \$image->xchg(0,1)
```

average processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## avgover

```  Synonym for average.
```

## daverage

```  Signature: (a(n); double [o]b())
```

Project via average to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the average along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = daverage(\$a);
```

``` \$spectrum = daverage \$image->xchg(0,1)
```

Unlike average, the calculation is performed in double precision.

daverage processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## davgover

```  Synonym for daverage.
```

## medover

```  Signature: (a(n); [o]b(); [t]tmp(n))
```

Project via median to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the median along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = medover(\$a);
```

``` \$spectrum = medover \$image->xchg(0,1)
```

medover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## oddmedover

```  Signature: (a(n); [o]b(); [t]tmp(n))
```

Project via oddmedian to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the oddmedian along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = oddmedover(\$a);
```

``` \$spectrum = oddmedover \$image->xchg(0,1)
```

The median is sometimes not a good choice as if the array has an even number of elements it lies half-way between the two middle values - thus it does not always correspond to a data value. The lower-odd median is just the lower of these two values and so it ALWAYS sits on an actual data value which is useful in some circumstances.

oddmedover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## modeover

```  Signature: (data(n); [o]out(); [t]sorted(n))
```

Project via mode to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the mode along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = modeover(\$a);
```

``` \$spectrum = modeover \$image->xchg(0,1)
```

The mode is the single element most frequently found in a discrete data set.

It only makes sense for integer data types, since floating-point types are demoted to integer before the mode is calculated.

"modeover" treats BAD the same as any other value: if BAD is the most common element, the returned value is also BAD.

modeover does not process bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## pctover

```  Signature: (a(n); p(); [o]b(); [t]tmp(n))
```

Project via percentile to N-1 dimensions

This function reduces the dimensionality of a piddle by one by finding the specified percentile (p) along the 1st dimension. The specified percentile must be between 0.0 and 1.0. When the specified percentile falls between data points, the result is interpolated. Values outside the allowed range are clipped to 0.0 or 1.0 respectively. The algorithm implemented here is based on the interpolation variant described at <http://en.wikipedia.org/wiki/Percentile> as used by Microsoft Excel and recommended by NIST.

By using xchg etc. it is possible to use any dimension.

``` \$b = pctover(\$a, \$p);
```

``` \$spectrum = pctover \$image->xchg(0,1), \$p
```

pctover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## oddpctover

```  Signature: (a(n); p(); [o]b(); [t]tmp(n))
```

Project via percentile to N-1 dimensions

This function reduces the dimensionality of a piddle by one by finding the specified percentile along the 1st dimension. The specified percentile must be between 0.0 and 1.0. When the specified percentile falls between two values, the nearest data value is the result. The algorithm implemented is from the textbook version described first at <http://en.wikipedia.org/wiki/Percentile>.

By using xchg etc. it is possible to use any dimension.

``` \$b = oddpctover(\$a, \$p);
```

``` \$spectrum = oddpctover \$image->xchg(0,1), \$p
```

oddpctover processes bad values. It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles.

## pct

Return the specified percentile of all elements in a piddle. The specified percentile (p) must be between 0.0 and 1.0. When the specified percentile falls between data points, the result is interpolated.

``` \$x = pct(\$data, \$pct);
```

## oddpct

Return the specified percentile of all elements in a piddle. The specified percentile must be between 0.0 and 1.0. When the specified percentile falls between two values, the nearest data value is the result.

``` \$x = oddpct(\$data, \$pct);
```

## avg

Return the average of all elements in a piddle.

``` \$x = avg(\$data);
```

## sum

Return the sum of all elements in a piddle.

``` \$x = sum(\$data);
```

## prod

Return the product of all elements in a piddle.

``` \$x = prod(\$data);
```

## davg

Return the average (in double precision) of all elements in a piddle.

``` \$x = davg(\$data);
```

## dsum

Return the sum (in double precision) of all elements in a piddle.

``` \$x = dsum(\$data);
```

## dprod

Return the product (in double precision) of all elements in a piddle.

``` \$x = dprod(\$data);
```

## zcheck

Return the check for zero of all elements in a piddle.

``` \$x = zcheck(\$data);
```

## and

Return the logical and of all elements in a piddle.

``` \$x = and(\$data);
```

## band

Return the bitwise and of all elements in a piddle.

``` \$x = band(\$data);
```

## or

Return the logical or of all elements in a piddle.

``` \$x = or(\$data);
```

## bor

Return the bitwise or of all elements in a piddle.

``` \$x = bor(\$data);
```

## min

Return the minimum of all elements in a piddle.

``` \$x = min(\$data);
```

## max

Return the maximum of all elements in a piddle.

``` \$x = max(\$data);
```

## median

Return the median of all elements in a piddle.

``` \$x = median(\$data);
```

## mode

Return the mode of all elements in a piddle.

``` \$x = mode(\$data);
```

## oddmedian

Return the oddmedian of all elements in a piddle.

``` \$x = oddmedian(\$data);
```

## any

Return true if any element in piddle set

Useful in conditional expressions:

``` if (any \$a>15) { print "some values are greater than 15\n" }
```

See or for comments on what happens when all elements in the check are bad.

## all

Return true if all elements in piddle set

Useful in conditional expressions:

``` if (all \$a>15) { print "all values are greater than 15\n" }
```

See and for comments on what happens when all elements in the check are bad.

## minmax

Returns an array with minimum and maximum values of a piddle.

``` (\$mn, \$mx) = minmax(\$pdl);
```

This routine does not thread over the dimensions of \$pdl; it returns the minimum and maximum values of the whole array. See minmaximum if this is not what is required. The two values are returned as Perl scalars similar to min/max.

``` pdl> \$x = pdl [1,-2,3,5,0]
pdl> (\$min, \$max) = minmax(\$x);
pdl> p "\$min \$max\n";
-2 5
```

## qsort

```  Signature: (a(n); [o]b(n))
```

Quicksort a vector into ascending order.

``` print qsort random(10);
```

Bad values are moved to the end of the array:

``` pdl> p \$b
pdl> p qsort(\$b)
```

## qsorti

```  Signature: (a(n); indx [o]indx(n))
```

Quicksort a vector and return index of elements in ascending order.

``` \$ix = qsorti \$a;
print \$a->index(\$ix); # Sorted list
```

Bad elements are moved to the end of the array:

``` pdl> p \$b
pdl> p \$b->index( qsorti(\$b) )
```

## qsortvec

```  Signature: (a(n,m); [o]b(n,m))
```

Sort a list of vectors lexicographically.

The 0th dimension of the source piddle is dimension in the vector; the 1st dimension is list order. Higher dimensions are threaded over.

``` print qsortvec pdl([[1,2],[0,500],[2,3],[4,2],[3,4],[3,5]]);
[
[  0 500]
[  1   2]
[  2   3]
[  3   4]
[  3   5]
[  4   2]
]
```

Vectors with bad components should be moved to the end of the array:

## qsortveci

```  Signature: (a(n,m); indx [o]indx(m))
```

Sort a list of vectors lexicographically, returning the indices of the sorted vectors rather than the sorted list itself.

As with "qsortvec", the input PDL should be an NxM array containing M separate N-dimensional vectors. The return value is an integer M-PDL containing the M-indices of original array rows, in sorted order.

As with "qsortvec", the zeroth element of the vectors runs slowest in the sorted list.

Additional dimensions are threaded over: each plane is sorted separately, so qsortveci may be thought of as a collapse operator of sorts (groan).

Vectors with bad components should be moved to the end of the array:

## minimum

```  Signature: (a(n); [o]c())
```

Project via minimum to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the minimum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = minimum(\$a);
```

``` \$spectrum = minimum \$image->xchg(0,1)
```

Output is set bad if all elements of the input are bad, otherwise the bad flag is cleared for the output piddle.

Note that "NaNs" are considered to be valid values; see isfinite and badmask for ways of masking NaNs.

## minimum_ind

```  Signature: (a(n); indx [o] c())
```

Like minimum but returns the index rather than the value

Output is set bad if all elements of the input are bad, otherwise the bad flag is cleared for the output piddle.

## minimum_n_ind

```  Signature: (a(n); indx [o]c(m))
```

Returns the index of "m" minimum elements

Not yet been converted to ignore bad values

## maximum

```  Signature: (a(n); [o]c())
```

Project via maximum to N-1 dimensions

This function reduces the dimensionality of a piddle by one by taking the maximum along the 1st dimension.

By using xchg etc. it is possible to use any dimension.

``` \$b = maximum(\$a);
```

``` \$spectrum = maximum \$image->xchg(0,1)
```

Output is set bad if all elements of the input are bad, otherwise the bad flag is cleared for the output piddle.

Note that "NaNs" are considered to be valid values; see isfinite and badmask for ways of masking NaNs.

## maximum_ind

```  Signature: (a(n); indx [o] c())
```

Like maximum but returns the index rather than the value

Output is set bad if all elements of the input are bad, otherwise the bad flag is cleared for the output piddle.

## maximum_n_ind

```  Signature: (a(n); indx [o]c(m))
```

Returns the index of "m" maximum elements

Not yet been converted to ignore bad values

## maxover

```  Synonym for maximum.
```

## maxover_ind

```  Synonym for maximum_ind.
```

## maxover_n_ind

```  Synonym for maximum_n_ind.
```

## minover

```  Synonym for minimum.
```

## minover_ind

```  Synonym for minimum_ind.
```

## minover_n_ind

```  Synonym for minimum_n_ind
```

## minmaximum

```  Signature: (a(n); [o]cmin(); [o] cmax(); indx [o]cmin_ind(); indx [o]cmax_ind())
```

Find minimum and maximum and their indices for a given piddle;

``` pdl> \$a=pdl [[-2,3,4],[1,0,3]]
pdl> (\$min, \$max, \$min_ind, \$max_ind)=minmaximum(\$a)
pdl> p \$min, \$max, \$min_ind, \$max_ind
[-2 0] [4 3] [0 1] [2 2]
```

```  Synonym for minmaximum.