The single-input version of the SAX2 function requires one input table, input.
The multiple-input version of the SAX2 function requires three input tables—input, meanstats, and stdevstats.
The input table must have one or more columns that contain time series data to be transformed, and you must specify their names with the ValueColumns argument. The input table can have other columns, but the function ignores them unless you specify them with the TimeColumn or Accumulate argument. The following table gives the valid data types for input table columns that you can specify with the ValueColumns, TimeColumn, and Accumulate arguments.
Column Name | Data Type | Description |
---|---|---|
value_column | BYTEINT, SMALLINT, INTEGER, BIGINT, NUMERIC, or DOUBLE PRECISION | Contains time series data to be transformed, specified with the ValueColumns argument. The input table must have at least one value_column. |
time_column | BYTEINT, SMALLINT, INTEGER, BIGINT, NUMERIC, or DOUBLE PRECISION | Contains the time axis of the data, specified with the TimeColumn argument. |
accumulate_column | Any | Column to copy to the output table. |
Both the meanstats and stdevstats tables must have every value_column and partition column (in key) that the input table has.
The meanstats table contains the global means of each value_column of the input table. Each row of the meanstats table specifies the global means for one input partition.
The stdevstats table contains the global standard deviations of each value_column of the input table. Each row of the stdevstats table specifies the global means for one input partition.