If you need to fill in missing values in a column with the last valid value from the rows above each missing value, you can use custom expression function LastValidBefore(). But this function uses the default row order from when the table is imported into Spotfire. If you need to do the calculation based on a certain custom order (based on a column sorting), this function does not give the desired results.

For example, you have a data table as the following:

A    B
1    0.3
2    
3    0.6
4    0.1
5    
6    
7    0.5
8    0.5

And you want to substitute the NULL values in column B with the last non-NULL value, based on the index in column A, to get a calculated column C, resulting in the desired outcome of:

A    B      C
1    0.3    0.3
2           0.3
3    0.6    0.6
4    0.1    0.1
5           0.1
6           0.1
7    0.5    0.5
8    0.5    0.5

​LastValidBefore([B])

A    B
1    0.3
2    
5    
6    
7    0.5
8    0.5
3    0.6
4    0.1

A    B      C
1    0.3    0.3
2           0.3 
5           0.3  
6           0.3  
7    0.5    0.5
8    0.5    0.5
3    0.6    0.6
4    0.1    0.1

To perform a similar calculation to LastValidBefore(), but based on a custom ordered index column (column A in our case) and not the default row order of the imported data table, then the expression you need to use should be:

LastValueForMax([A],[B]) OVER (AllPrevious([A]))

A    B      C	
1    0.3    0.3
2           0.3 
5           0.1  
6           0.1  
7    0.5    0.5
8    0.5    0.5
3    0.6    0.6
4    0.1    0.1

A    B      C
1    0.3    0.3
2           0.3
3    0.6    0.6
4    0.1    0.1
5           0.1
6           0.1
7    0.5    0.5
8    0.5    0.5

• https://docs.tibco.com/pub/sfire-analyst/7.11.0/doc/html/en-US/TIB_sfire-analyst_UsersGuide/ncfe/ncfe_statistical_functions.htm