The DWT2D function has a required input table (or view) and an optional wavelet filter table. If you omit the wavelet filter table, you must specify the Wavelet argument.
The input table can contain at most 1594 columns. The function assumes that each sequence can be fitted into the memory of the worker. The following table describes the input table columns that you can or must specify with arguments. The input table can have additional columns, but the function ignores them.
The table below shows the schema of the wavelet filter table.
Column Name | Data Type | Description |
---|---|---|
partition_column | INTEGER, SMALLINT, BIGINT, NUMERIC [(p, s)], TEXT, VARCHAR, VARCHAR(n), UUID, or BYTEA | Optional. Sequence identifier of the sequence to which the data belongs. Rows with the same partition column values belong to the same sequence. If the input table has multiple partition columns, the function treats the first one as the distribute key of the output and meta tables. If all the values are invalid in a sequence, the function ignores the sequence. |
indexy | INTEGER | Y index of the sequence to which the data belongs. If this column contains a NULL value, the function ignores the row. |
indexx | INTEGER | X index of the sequence to which the data belongs. If this column contains a NULL value, the function ignores the row. |
input_column | INTEGER, SMALLINT, BIGINT, BIGSERIAL, DOUBLE PRECISION, NUMERIC [(p, s)], or SERIAL | Data to be transformed—numeric values between -1e308 and 1e308 (or NULL, which the function treats as 0). |