currently I've implemented a new algorithm where only an upper bound for the column count of the output table is known before executing the node. When the upper bound is e.g. 100 and the algorithm only returns e.g. 90 columns Knime throws an error because it expected 100 columns as it is defined in the configure method.
Is there an easy way (and code samples) to create an output where the number of output columns is only known after algorithm/node execution?