Ophidia Primitives Manual

The links below describe the set of array-based primitives available in the platform. Currently available array-based functions allow data sub-setting, data aggregation (i.e. max, min, avg), array concatenation, algebraic expressions and predicate evaluation. Core functions of well-known numerical libraries (e.g. GSL) have been included into the primitives.

Each manual page describes the primitive’s functionalities, the input arguments required, the returned type and a simple example. To uniform the interfaces, almost all primitives use as first two arguments input and output measure type, even when these parameters are not necessary. In these cases they are marked as not used.

Most operators rely on primitives to perform array-based operations, however to execute a selected primitive, a special operator OPH_APPLY must be used.

Ophidia primitives have been developed as MySQL User Defined Functions (UDF), see development guide for more information, hence the functions can be also used in a nested fashion.

Core Array

NAME DESCRIPTION
oph_append It concats multiple input measures into a single output measure.
oph_concat It builds a new measure array concatenating the measures specified.
oph_concat2 It builds a new measure array concatenating the measures specified; the primitive appends one array per row by selecting it cyclically from the set.
oph_count_array It counts the number of elements into an array.
oph_expand It expands an array by putting NaN in given positions.
oph_extend It creates an array by concating more copies of input array.
oph_find It finds the number of occurences into a measure array that are inside the interval [value-distance;value+distance].
oph_gsl_sort It orders the elements of the measure array in an ascending way using heapsort.
oph_interlace It interlaces multiple input measures into a single output measure.
oph_padding It appends padding values so that array size becomes a given value.
oph_permute It executes a permutation of the implicit dimension order.
oph_reverse It extracts a binary array of measures in a reversed order (from the last to the first).
oph_roll_up It transforms an explicit dimension in implicit.
oph_rotate It rotates the values inside a multidimensional measure array.
oph_size_array It computes the size of an array (length of the array in bytes).
oph_shift It shifts the values inside a measure array by ‘offset’ positions using the ‘filling’ value.

Selection

NAME DESCRIPTION
oph_extract It returns a subset of a measure obtained using the (long long) indexes passed as ‘index. list’.
oph_get_subarray2 It extracts a subset of a measure array. The subset cannot be applied on multiple dimensions.
oph_get_subarray3 It extracts a subset of a measure array. The subset can also be applied on multiple dimensions.
oph_get_subarray It extracts a subset of ‘count’ elements of an array starting from ‘start’.
oph_mask_array It applies a mask to a measure array.

Arithmetic

NAME DESCRIPTION
oph_abs_array It computes the modulus of each value of measure array b and the value in the same position in measure array a.
oph_arg_array It computes the phase of each value of measure array b and the value in the same position in measure array a.
oph_compare It returns 0 if the two input measures are equal.
oph_div_array It divides each value of measure array a for the value in the same position in measure array b.
oph_gsl_complex_get_abs It computes the absolute values of a binary array of complex numbers.
oph_gsl_complex_get_arg It computes the argument (phase) of a binary array of complex numbers.
oph_gsl_complex_get_imag It extracts the imaginary parts from a binary array of complex numbers.
oph_gsl_complex_get_real It extracts the real parts from a binary array of complex numbers.
oph_gsl_complex_to_polar It converts a binary array of complex numbers expressed in cartesian coordinates into the polar form.
oph_gsl_complex_to_rect It converts a binary array of complex numbers expressed in polar form into the rectangular form.
oph_math It executes a scalar mathematical operation on each element of the measure array.
oph_max_array It extracts the maximum between the values in the same position in measure array a and array b.
oph_min_array It extracts the minimum between the values in the same position in measure array a and array b.
oph_mul_array It multiplies each element of the measure array by the scalar value.
oph_mul_scalar It multiplies each element of the measure array by the scalar value.
oph_mul_scalar2 It multiplies each element of the measure array by a scalar value. It can operate also with complex numbers.
oph_operator_array It applies an operation between the value of measure array b and the value in the same position in measure array a.
oph_operator It applies an operator on the whole measure array.
oph_sum_array It sums each value of measure array b to the value in the same position in measure array a.
oph_sum_scalar It sums the scalar value to each element of the measure array.
oph_sum_scalar2 It sums the scalar value to each element of the measure array. It can operate also with complex numbers.
oph_sub_array It subtracts each value of measure array b from the value in the same position in measure array a.

Statistical

NAME DESCRIPTION
oph_aggregate_operator It aggregates different measure arrays performing the specified operation on each column of array elements.
oph_aggregate_stats_final It computes a statistical index working vertically on columns of arrays of partial results obtained from oph_aggregate_stats_partial.
oph_aggregate_stats_partial It computes partial results useful to compute a statistical index value working vertically on the columns of arrays of measures.
oph_aggregate_stats It computes a statistical index based on mask value working vertically on the columns of arrays of measures.
oph_gsl_boxplot It computes an array with the five indicators used for the boxplot diagram.
oph_gsl_correlation It computes the Pearson coefficient of two measure arrays.
oph_gsl_histogram It estimates the empiric distribution of a sample of measures contained in the array with n bins.
oph_gsl_quantile It computes the quantile(s) specified as q1,q2,... on the elements of the measure array ordered in an ascending way.
oph_gsl_sd It computes the standard deviation of a binary array.
oph_gsl_stats It computes statistical indexes of the measure array. The type of indexes depends on a mask.
oph_normalize It normalizes a measure array performing the specified operation on ‘count’ elements.
oph_quantize It quantizes an array according to a binary list of quantization level and a binary list of quantization values (expressed as doubles).
oph_reduce It reduces a measure array performing the specified operation on ‘count’ elements.
oph_reduce2 It reduces a measure array performing the specified operation on blocks of count elements. It supports multiple implicit dimension.
oph_reduce3 It reduces a measure array performing the specified operation on blocks of elements (aggregation sets).

Transformation

NAME DESCRIPTION
oph_accumulate It returns an array in which the ‘i’-th element is the cumulative sum of the first ‘i’ elements of input array.
oph_deaccumulate It returns an array in which the ‘i’-th element is the difference between the ‘i’-th and ‘i-1’-th element of input array.
oph_gsl_dwt It computes the Discrete Wavelet Transform of a measure array.
oph_gsl_fft It computes the Fast Fourier Transform of a measure array.
oph_gsl_idwt It compute the Inverse Discrete Wavelet Transform of a measure array.
oph_gsl_ifft It computes the inverse Fast Fourier Transform of a measure array.
oph_moving_avg It computes the moving average operation on each element of a measure array.
oph_predicate It evaluates a predicate.
oph_predicate2 It evaluates a predicate.
oph_replace It replacea items of a measure array with other values or suppress them.

Numerical Analysis

NAME DESCRIPTION
oph_gsl_fit_linear_coeff It computes the linear regression coefficients.
oph_gsl_fit_linear It computes the linear regression.
oph_gsl_spline It computes the interpolation based on cubic splines.

Mining

NAME DESCRIPTION
oph_ccluster_kcluster It computes clusters from input data through the use of k-means or k-medians algorithms.

Miscellaneous

NAME DESCRIPTION
oph_cast It converts the measure array value type from input to output type.
oph_convert_d It converts a binary element in a real value.
oph_convert_l It converts a binary element in an integer value.
oph_dump It translates a binary array into a human-readable form.
oph_get_index_array It returns a binary array with values from ‘start value’ from to ‘stop value’.
oph_to_bin It converts a string in a form ‘n1, n2, n3, ...’ into a binary array of measures.
oph_value_to_bin It converts a value into a binary array of measures.