loop end int to double conversion

Hi,

I have a Loop in which I read in a table in each iteration and use the Domain Calculator and Extraxt Table Spec nodes. This loop is inside a parallel chunk loop. When creating the workflow I have to work on the first row of the first chunk. The upper and lower bounds are of type double.

Inside the chunk loop one iteratation is starting with a case where the Extract Table Specs is creating an Int Cell for the upper and lower bounds. This in the end causes a problem because the next file in the same chunk iteration is of type DoubleCell, which is creating a problem. (

ERROR	 Loop End                      	 Execute failed: Input table's structure differs from reference (first iteration) table: Column 3 [Lower Bound (IntCell)] vs. [Lower Bound (DoubleCell)]Column 4 [Upper Bound (IntCell)] vs. [Upper Bound (DoubleCell)]

)

I tried to use the Column Rename node but this is not too easy because the firt iteration of the first chunk produces a DoubleCell, thus I cannot really change the column properties. I thought of defining a variable and forcing a change to Double but that seems rather complicated. Do you have any other suggestions?

Thanks,

Bernd

One solution to this type of situation is to create an empty table structure using Table Creator node for example with the correct column names, and make the columns to be of the type you wish, I.e. Double. You now concatenate this empty table in to your data test inside the loop. This will force the data set to always be double in each loop iteration.

simon.

Hi Bernd,

can you post me the workflow?

Iris

Or you could try a Math node to return 1.0 * yourColumn - the .0 is important, as it will force the result to be a double

Steve

Thanks for the relpies. I have adopted Simon's suggestion and it works. In the end I actually also included Group Loop start to be able to distribute the load better, but that is rather a problem with chunk loop ...

Thanks a lot!

Bernd