DWT Syntax Elements - Teradata Vantage

Machine Learning Engine Analytic Function Reference

Product
Teradata Vantage
Release Number
9.02
9.01
2.0
1.3
Published
February 2022
Language
English (United States)
Last Update
2022-02-10
dita:mapPath
rnn1580259159235.ditamap
dita:ditavalPath
ybt1582220416951.ditaval
dita:id
B700-4003
lifecycle
previous
Product Category
Teradata Vantageā„¢
OutputTable
Specify the name for the table that the function creates to store the coefficients output by the wavelet transform. This table must not exist.
MetaInformationTable
Specify the name for the table that the function creates to store the meta information for the wavelet transform. This table must not exist.
TargetColumns
Specify the names of the InputTable columns that contain the data to transform. These columns must contain numeric values between -1e308 and 1e308. The function treats NULL as 0.
SortColumn
Specify the name of the InputTable column that defines the order of samples in the sequences to transform. In a time series sequence, the column can consist of timestamp values.
If sort_column has duplicate elements in a sequence (that is, in a partition), sequence order can vary, and the function can produce different transform results for the sequence.
PartitionColumns
[Optional] Specify the names of the InputTable partition columns, which identify the sequences. Rows with the same partition column values belong to the same sequence. If you specify multiple partition columns, the function treats the first one as the distribute key of OutputTable and MetaInformationTable.
The DWT output tables are the IDWT input tables. If you specify this syntax element for DWT, you must also specify it for IDWT; otherwise, the results might not make sense.
Default behavior: All rows belong to one sequence, and the function creates a distribute key column named dwt_id_random_name in both the OutputTable and MetaInformationTable. In both tables, every cell of dwt_id_random_name has the value 1.
Wavelet
[Required if you omit WaveletFilterTable.] Specify a wavelet filter name from the following table.
Wavelet Family Supported Wavelet Names (wavelet values)
Daubechies 'db1' or 'haar', 'db2', .... ,'db10'
Coiflets 'coif1', ... , 'coif5'
Symlets 'sym1', ... ,' sym10'
Discrete Meyer 'dmey'
Biorthogonal 'bior1.1', 'bior1.3', 'bior1.5', 'bior2.2', 'bior2.4', 'bior2.6', 'bior2.8', 'bior3.1', 'bior3.3', 'bior3.5', 'bior3.7', 'bior3.9', 'bior4.4', 'bior5.5'
Reverse Biorthogonal 'rbio1.1', 'rbio1.3', 'rbio1.5' 'rbio2.2', 'rbio2.4', 'rbio2.6', 'rbio2.8', 'rbio3.1', 'rbio3.3', 'rbio3.5', 'rbio3.7','rbio3.9', 'rbio4.4', 'rbio5.5'
WaveletTransformLevel
Specify the wavelet transform level. The value level must be an integer in the range [1, 1000].
ExtensionMode
[Optional] Specify the wavelet transform level. The level must be an integer in the range [1, 1000].

For the examples in the following table, assume that the sequence before the extension is 1 2 3 4 and the convolution kernel in the wavelet filter has the length 6, which means that the length of the sequence is to extend by 5 positions before and after the sequence.

Supported Extension Mode (extension_mode value) Description
sym (Default) Symmetrically replicate boundary values, mirroring the points near the boundaries. For example:

4 4 3 2 1 | 1 2 3 4 | 4 3 2 1 1

zpd Zero-pad boundary values with zero. For example:

0 0 0 0 0 | 1 2 3 4 | 0 0 0 0 0

ppd Periodic extension, fill boundary values as the input sequence is a periodic one. For example:

4 1 2 3 4 | 1 2 3 4 | 1 2 3 4 1