Filter rows based on numerical/expression column

Author

Cox Lab

Published

June 27, 2024

1 General

2 Brief description

Only those rows are kept that have a value in the numerical column fulfilling the equation or inequality relation.

Output: The filtered matrix.

3 Parameters

3.1 Number of columns

The filtering is based on relations of expression/numerical columns. Up to five columns (default: 1) can be selected. Depending on the number of chosen columns drop down box(es) appear on the pop-up window named “x”, “y”, “z”, “a” and “b”. In these drop down boxes expression/numerical columns can be specified, which can then be used in the relations that should be applied for filtering the matrix.

3.2 Number of relations

Up to five relations using the previously specified columns can be included in the filtering process (default:1). Depending on the selected number of relations text fields on the pop-up window appear named “Relation 1”, “Relation 2”, “Relation 3”, “Relation 4” and “Relation 5”.

In each text field a relation for the filtering process can be defined using the variables of the parameter “Number of columns”. For the relations numbers with “.” as decimal point, “+”, “-”, “*”, “/” and “^” as well as scientific notation (e.g. “5.4e-12”) can be used.

3.3 Combine through

Defines how the specified relations are combined (default: intersection). Depending on the specified combination mode either rows, which fulfill the “intersection” (default) of the relations are kept or the ones fulfilling the “union”.

3.4 Filter mode

The “Filter mode” defines, whether the input matrix will be reduced (“Reduce matrix” = default) or a new categorical column called “Filter” will be generated containing the categories “Keep” and “Discard” (“Filter mode” = “Add categorical column”).

4 Parameter window