Join

) dataset to combine the solutions from two datasets. The Join dataset makes it possible to compare solutions from two datasets — for example, to evaluate and visualize the difference between two solutions to the same problem using two different meshes in a mesh convergence study or to create the sum over a parametric sweep that contains a few solutions. The Join dataset has predefined methods to combine the solutions to get the difference, norm of difference, product, quotient, or sum of the two solution datasets. In addition, two predefined “operators”, data1 and data2, correspond to the solution data in the first and second Solution dataset, respectively, and make it possible to compare, for example, solutions from different but compatible models or to combine the two solution datasets using a method other than the ones that you can select directly.


	You can also use the withsol operator in a plot expression, for example, to combine solutions from different solver sequences. See withsol for more information.


	When combining data from two meshes, the Join dataset searches for the points of one mesh in the other; the results are then combined. Note that this process is not symmetric.

Right-click a dataset node to add a subnode if desired and if its two source datasets use the same geometry.

Data 1 and Data 2

Select a dataset as data1 and another dataset as data2 from the lists in the and sections, respectively.

Select which solutions to use from the lists:

•

Select (the default) to use all solutions in the dataset.

•

Select to use one of the available solutions in a time-dependent, parametric, or eigenvalue solution from the list of solutions that appear underneath the list.

The following dataset combinations support the option to include all solutions from both datasets:

•

Both datasets point to the same solution.

•

Both datasets point to two different stationary solutions.

•

Both datasets point to two different time-dependent solutions.

Combination

In the list, select a method for combining the datasets:

•

Select (the default) to combine the datasets using a difference: data1-data2.

•

Select to combine the datasets as abs(data1-data2), where data1 and data2 are the results of evaluating the expression in the two source datasets. For complex-valued data, this corresponds to the Euclidean norm of the difference.

•

Select to combine the datasets using an explicit expression with the “operators” data1 and data2 in, for example, a plot node’s field. This can be useful to compare two different dependent variables in two different datasets such as two solutions from two different models using the same geometry. For example, to plot the sum of the variable u from the first dataset and the variable v from the second dataset, times two, type 2*(data1(u)+data2(v)). The scope for data1 is the model to which the solution dataset under belongs, and similarly for data2.

•

Select to combine the datasets using a general expression in data1 and data2 that you type into the field. The default, data1-data2, is identical to the method. This method is useful for combining datasets with similar solution data (from a mesh convergence study, for example) using a method other than a difference, product, quotient, or sum.

•

Select to combine the datasets using a product: data1*data2.

•

Select to combine the datasets using a quotient: data1/data2.

•

Select to combine the datasets using a sum: data1+data2.


	If you choose Difference, for example, an operation such as sin(u) for the Join dataset is equivalent to sin(data1(u))-sin(data2(u)). If you want to evaluate or plot sin(u) as the operator applied to the difference itself, for example, choose Explicit and use sin(data1(u)-data2(u)) as the expression to evaluate or plot.


	It can happen that the geometries in the two datasets do not match. In such cases, for example, Join datasets that contain data from a domain cannot be selected in a point plot. Also, selections are not supported when using Join datasets.


	If you only want to compare the values of integrals over datasets, then you can achieve this using the Transformation type set to Difference in an Evaluation Group node, which provides better accuracy.