I think it's doable using UFL, if the number of columns has some practical limit. If we take 100 columns, define the columns as valuecolumn001 - valuecolumn100.
First read all tagnames when reading the header line, and then for each line containing data, do 100 writeToPi calls, one for each column.
- Use some C# code do to preprocessing before using UFL. Should be easy to code.
- Use SQL Server Integration Services and dynamic pivots. Very heavyweight option though, used it before, but would not quickly recommend if you have other options.
Hope this helps!
We'll stick with UFL then and just do multiple variables for the maximum possible number of columns that could be in the data file. Just wanted to make sure we weren't missing a feature/possibility with UFL
Thank you for your help and advice. Just as Max says, we didn't miss a feature of UFL.